Recomendación para La Evaluacion de de Cultivo Celulares y Sustratos para El Proceso de Manufactura de Biologicos

Annex 3
Recommendations for the evaluation of animal cell cultures

as substrates for the manufacture of biological medicinal
products and for the characterization of cell banks
Replacement of Annex 1 of WHO Technical Report Series, No. 878
Abbreviations 81
1. Introduction 84
2. Historical overview 84
3. Scope 88
4. Definitions 89
5. General considerations 95
5.1 Types of animal cell substrates 95
5.2 Potential risks and risk mitigation associated with biologicals produced in
animal cell cultures 100
Part A. General recommendations applicable to all types of cell
culture production 107
A.1 Good manufacturing practices 107
A.2 Principles of good cell culture practice 107
A.3 Selection of source materials 112
A.4 Certification of cell banks by the manufacturer 118
A.5 Cryopreservation and cell banking 120
Part B. Recommendations for the characterization of cell banks of
animal cell substrates 123
B.1 General considerations 123
B.2 Identity 126
B.3 Stability 128
B.4 Sterility 130
B.5 Viability 130
B.6 Growth characteristics 130
B.7 Homogeneity 131
B.8 Tumorigenicity 131
B.9 Oncogenicity 139
B.10 Cytogenetics 141
B.11 Microbial agents 142
B.12 Summary of tests for the evaluation and characterization of animal cell substrates 165
79
WHO Expert Committee on Biological Standardization Sixty-first report
Authors 166
Acknowledgements 168
References 168
Appendix 1
Tests for bovine viruses in serum used to produce cell banks 176
Appendix 2
Tumorigenicity protocol using athymic nude mice to assess mammalian cells 180
Appendix 3
Oncogenicity protocol for the evaluation of cellular DNA and cell lysates 184
Recommendations published by WHO are intended to be scientific

and advisory. Each of the following sections constitutes guidance
for national regulatory authorities (NRAs) and for manufacturers of
biological products. If an NRA so desires, these recommendations
may be adopted as definitive national requirements, or modifications
may be justified and made by the NRA. It is recommended that
modifications to these recommendations be made only on condition
that the modifications ensure that the biological product is at least
as safe and efficacious as that prepared in accordance with the
recommendations set out below. The parts of each section printed
in small type are comments for additional guidance, intended for
manufacturers and NRAs, which may benefit from these details.
WHO Technical Report Series No. 978, 2013
80
Annex 3
Abbreviations
ALS antilymphocyte serum
ATCC American Type Culture Collection
ATG antithymocyte globulin
ATS antithymocyte serum
BAV5 bovine adenovirus 5
BCG bacille Calmette–Guérin vaccine
bp base pairs
BPIV3 bovine parainfluenza type 3 virus
BPV bovine parvovirus
BRSV bovine respiratory syncytial virus
BSE bovine spongiform encephalopathy
BTV bluetongue virus
BVDV bovine viral diarrhoea virus
CCL continuous cell line
CEF chick embryo fibroblasts
CHO Chinese hamster ovary
CJD Creutzfeldt–Jakob disease
CPE cytopathic effect
CTL cytotoxic T-lymphocyte
CWD chronic wasting disease
DCL diploid cell line
EBV Epstein–Barr virus
ECB extended cell bank
EFSA European Food Safety Authority
ELISA enzyme-linked immunosorbent assay
EMA European Medicines Agency
EOP end of production
EOPC end-of-production cell
81
EPIC-PCR exon-primed intron crossing PCR

FBS fetal bovine serum
FDA Food and Drug Administration (USA)
FFI fatal familial insomnia
GBR geographical BSE-risk level
GMP good manufacturing practices
GSS Gerstmann–Sträussler–Scheinker syndrome
HCP hamster cheek pouch
HDC human diploid cell
HEK human embryonic kidney
HIV human immunodeficiency virus
HLA human leukocyte antigen
IBR infectious bovine rhinotracheitis virus
ICH International Conference on Harmonisation of Technical
Requirements for Registration of Pharmaceuticals for Human
Use
IFA immunofluorescence assays
IFN interferon
LCMV lymphocytic choriomeningitis virus
MAbs monoclonal antibodies
MCB master cell bank
MDCK Madin–Darby canine kidney

mRNA messenger RNA
miRNA microRNA
MPS massively parallel sequencing
NAT nucleic acid amplification technique
NCL national control laboratory
NK natural killer (cell)
NOD non-obese diabetic
NRA national regulatory authority
82
Annex 3
OIE World Organisation for Animal Health

PBS phosphate-buffered saline
PCC primary cell culture
PCR polymerase chain reaction
PDL population doubling level
PERT product-enhanced reverse transcriptase
PMCA protein misfolding cyclic amplification
PrP prion protein
RCB reference cell bank
rcDNA residual cellular DNA
rDNA recombinant DNA
REO3 reovirus 3
RFLP restriction fragment length polymorphism
RT reverse transcriptase
SCID severe combined immunodeficiency
SCL stem cell line
SPF specific-pathogen-free
STR short tandem repeats
SV simian virus
TCID50 median tissue culture infective dose
TEM transmission electron microscopy
TPD50 tumour-producing dose at the 50% end-point
TSE transmissible spongiform encephalopathy
vCJD variant CJD
VNTR variable number of tandem repeats
VSV vesicular stomatitis virus
WCB working cell bank
83
1. Introduction
Cell substrates are cells used to manufacture biological products. It is well
established that both cell substrates themselves and events linked to cell growth
can affect the characteristics and safety of the resultant biological products.
Therefore, a thorough understanding of the characteristics of the cell substrate is
essential in order to identify points of concern and to develop a quality control
system that addresses these points.
Recent advances in the use and quality control of new animal cell
substrates – particularly continuous cell lines (CCLs) and insect cells – led to the
conclusion that an update to the WHO requirements (Requirements for the use
of animal cells as in vitro substrates for the production of biologicals) (1) should
be prepared. In order to facilitate the resolution of regulatory/scientific issues
related to the use of animal (including human) cell cultures as substrates for the
production of biological products, WHO initiated this revision of its requirements
on cell substrates by establishing a WHO Study Group on Cell Substrates. Animal
cells refer to cells derived from organisms classified as within the animal kingdom.
This document is the result of the Study Group’s work, which included a wide
range of consultations with individuals and organizations with expertise in this
area. After comments were received from this consultative process, and from
invited reviewers, further revision of the draft recommendations was undertaken
and presented to the WHO Expert Committee on Biological Standardization in
2010. During the development of this document, guidance on the topic issued
by other relevant organizations was considered. An effort was made to make the
recommendations compatible with existing guidance whenever possible.
These Recommendations provide guidance to national regulatory
authorities (NRAs), national control laboratories (NCLs) and manufacturers on
basic principles and, in some cases, on detailed procedures that it is appropriate
to consider in the characterization of animal cells proposed for use in the
manufacture of biological products. Although the decision-making authority lies

with the NRA, it is advisable that NCL experts on this topic should be consulted.
2. Historical overview
Historically, the major concerns regarding the safety of biological medicinal
products manufactured in animal cells have been related to the possible presence
of microbial contaminants and, in some cases, to the properties and components
of the cells themselves – such as DNA and proteins.
For instance, in 1954 an experimental adenovirus vaccine was being
developed and human tumour cells (HeLa) were rejected as the cell substrate in
favour of “normal” cells (2). At that time, relatively little was known about the
84
Annex 3
biological mechanism(s) that lead to human cancer, so the risks to the recipients
of a vaccine based on HeLa cells could not be assessed and quantified scientifically.
Although “normal” cells were not defined, that decision led to the use of primary
cell cultures (PCCs) from animals such as monkeys, hamsters and embryonated
eggs for vaccine research and development (3).
The first Requirements for cell substrates were published by WHO in
1959 and related to the production of inactivated poliomyelitis vaccine in PCCs
derived from the kidneys of clinically healthy monkeys (4). Those Requirements
were revised and published in 1966 (5). Subsequently, other PCCs were used for
the production of other viral vaccines.
In the 1960s, human diploid cells (HDCs) were developed and proposed
as an alternative to primary monkey kidney cell cultures for production of polio
virus vaccine, as well as for production of other viral vaccines. The rationale for
using HDCs was based on the ability to:
■ cryogenically preserve the cells at low population doubling levels
(PDLs);
■ establish and characterize cryopreserved banks of cells that later
could be expanded to provide a standardized source of cells for
many decades;
■ extensively test recovered cells before use in vaccine production;
■ demonstrate that the cells were free from detectable adventitious
agents and that they were unable to form tumours when inoculated
into immunosuppressed animals.
Thus, HDCs were normal by all of the then existing criteria. It was argued
that because HDCs were normal and could be standardized, tested and used for
many years, they were a significant improvement over PCCs.
The path to acceptance of HDCs was long and difficult, primarily because
some members of the scientific community believed that HDCs might contain
a latent and unknown human oncogenic agent and that such a theoretical
agent posed a risk to the recipients of vaccines produced in HDCs. Numerous
conferences and discussions of new data eventually led to the acceptance of HDCs
as a substrate for viral vaccine production, and they continue to be used by many
manufacturers for various viral vaccines that have a long history of safety and
effectiveness. The concept of a master cell bank (MCB) and working cell bank
(WCB) system and the characterization of the cell substrate were introduced
during that period (6, 7).
Both the understanding of tumour cell biology and the technological
tools that were available at that time were much more limited than they are today.
As a result, the proponents of using HDCs for vaccine production based their
85
argument that the cells were normal, and therefore safe to use, on four points,
namely: freedom from detectable adventitious agents; the finite life of HDCs; the
diploid nature of HDCs; and the inability of HDCs to form tumours in various
in vivo test systems.
In order to provide a high level of assurance that those four characteristics
were stable, the initial lot release tests for each batch of a vaccine derived from
HDCs included tests of the cell substrate for adventitious agents, karyology
and tumorigenicity (8, 9). The main question that was being addressed by the
routine use of tumorigenicity tests was whether or not the production cell
culture had undergone a contamination or transformation event such that it
contained a mixture of “normal” and tumorigenic cells. It was eventually agreed
that tumorigenicity testing was not sufficiently sensitive to detect a low level
of tumorigenic cells, and that it was a waste of animals and time to carry out
repeated testing of a cell line that had been well characterized and would be used
in the context of a cell bank system. Thus, tumorigenicity tests were eventually
required only for the characterization of an MCB (using cells at the proposed in
vitro cell age for production or beyond) for both HDCs and CCLs (10, 11).
In the 1970s, there was a need in clinical research for more interferon
alpha (IFN-α) than could be produced from primary human lymphocytes. In
response, human tumour cells (Namalwa) grown in vitro were proposed as
a cell substrate for the production of IFN-α. The primary concerns about the
use of Namalwa cells were that they contained the Epstein–Barr virus (EBV)
genome integrated into the cellular DNA, and that either whole virus or DNA
containing viral elements could be transmitted to the recipients of the IFN-α
product. Nevertheless, by the end of the 1970s, regulatory agencies had allowed
human clinical studies to commence, and the product was eventually approved
in several countries. Among the most important factors contributing to those
decisions was the fact that IFN, as opposed to live viral vaccines, was not a
replicating agent, and IFN-α was being used as a therapeutic product rather than
a prophylactic one, thus representing different risk–benefit considerations. In

addition, technology had advanced significantly so that IFN-α could be highly
purified and the purification process could be validated to demonstrate that EBV
and cellular DNA were undetectable in the final product, within the limits of the
assays then available, which permitted risk mitigation.
In the 1980s, advances in science and technology led to the development
of recombinant DNA (rDNA)-derived proteins and monoclonal antibodies
(MAbs). Animal cells with the capacity to grow continuously in vitro (CCLs)
were the substrates of choice for those products because of the ease with which
they could be transfected and engineered. Also, in contrast to PCCs and HDCs,
CCLs grew rapidly to achieve a high density and expressed a variety of products
at high concentrations. Chinese hamster ovary (CHO) cell lines became widely
86
Annex 3
used for rDNA products, and hybridomas of various types were required for
the production of MAbs. The use of such cells as substrates in the manufacture
of a large array of potentially important biological medicinal products raised
safety concerns once again. A scientific consensus emerged from numerous
conferences that there are three major elements of potential concern related to
animal-cell substrates – DNA, viruses and transforming proteins. In 1986, WHO
established a WHO Study Group on Cell Substrates to examine cell substrate
issues in greater depth.
The Study Group concluded that there is no reason to exclude CCLs from
consideration as substrates for the production of biologicals, and that CCLs are
in general acceptable when the manufacturing process is shown to eliminate
potential contaminating viruses that are pathogenic for humans and to reduce
DNA to acceptable levels and/or eliminate its biological activity (12). The Study
Group’s emphasis on infectious agents as the major risk factor was based largely
on experience of virus transmission and disease occurring through contaminated
biological products (e.g. hepatitis B virus and HIV in factor VIII). WHO’s
Requirements for CCLs used for the production of biologicals were published in
1987 (13). On the basis of a review of more recent data, those Requirements were
revised in 1998 to raise the acceptable level of rcDNA to 10 ng per parenteral
dose. In addition, it was pointed out that beta-propiolactone, a viral inactivating
agent, may also destroy the biological activity of DNA. Use of this agent therefore
provides an additional level of confidence, even when the amount of DNA per
dose may be substantial (1).
During the 1990s and into the 2000s, a variety of CCLs were explored
as cell substrates for biological products in development because, like the cell
lines referred to above, they offered significant advantages during production
(e.g. rapid growth and high expression). They include the tumorigenic cell lines –
HeLa for adeno-associated virus vectored HIV vaccines, PER.C6 for influenza
and HIV vaccines, Madin–Darby canine kidney (MDCK) for influenza vaccines,
and 293ORF6 for HIV vaccines. More recently, insect cell lines and stem cell
lines (SCLs) have been proposed for the manufacture of biological products, and
such cells introduce a new set of challenges with regard to their evaluation and
characterization.
The acceptability of a given cell type (primary, diploid, stem or continuous)
as a substrate for the production of a specific biological product depends on a
variety of factors, including in-depth knowledge of the cell type’s basic biological
characteristics. It is important to recognize that the tumorigenic potential of a
CCL is only one of many factors, including the extent to which the manufacturing
process reduces or eliminates cellular factors that may be of concern, to be
considered. An assessment of the totality of the data available is needed in
order to determine whether a product manufactured in a given cell substrate is
potentially approvable.
87
The following Recommendations provide guidance to manufacturers,

NRAs and NCLs on the evaluation of animal cell cultures used as substrates for
the production of biological medicinal products, and for the characterization of
cell banks.
The main changes compared to the Requirements published in WHO
Technical Report Series, No. 878, Annex 1, include the following:
■ general manufacturing recommendations applicable to all types of
cell culture production have been updated;
■ some considerations for evaluating new cell substrates such as insect
cells and stem cells have been added;
■ definitions have been updated and expanded in number and scope,
and are moved to an earlier point in the document;
■ the structure of the document has been modified to include more
background information, and the applicability of various sections to
different types of cell substrates is highlighted;
■ a new section on risk-reduction strategies during the manufacture
of biological products has been added;
■ a section on good cell culture practice has been added;
■ the section on selection of source materials has been updated, and
the detailed methods used to test for bovine viruses in serum have
been added in Appendix 1;
■ tumorigenicity testing has been updated, and a model protocol for
the nude mouse model has been added in Appendix 2;
■ oncogenicity testing of tumorigenic cell lysates has been added, and
a model protocol is added in Appendix 3;
■ recommendations for acceptable levels of residual cellular DNA are
product specific and are not specifically addressed;

■ recommendations for microbial agents testing have been updated.
3. Scope
These Recommendations supersede previous WHO Requirements or
Recommendations describing procedures for the use of animal cell substrates for
the production of biological medicinal products (1, 13).
Some of the recommendations may also be useful in the quality control
of specific biological products during the manufacturing process, but it is beyond
the scope of this document to recommend quality control release tests. Likewise,
risk-based assessments related to product approvals are beyond the scope of this
88
Annex 3
document. Requirements or recommendations for individual products should be

consulted for such assessments.
Cells modified by recombinant DNA technology have been increasingly
used in the manufacture of novel medicinal products, and specific considerations
for those products are addressed elsewhere (1, 10, 14, 15). Nevertheless, a number
of generic issues apply to genetically modified and other cell substrates.
These Recommendations specifically exclude all products manufactured
in embryonated eggs, microbial cells (i.e. bacteria and yeast) and plant cells. Also
excluded are whole, viable animal cells such as stem cells when they are used
directly for therapy by transplantation into patients or when they are developed
into SCLs for the purpose of using them as therapeutic agents by transplantation.
In those cases, characterization tests should be discussed with the NRA/NCL.
Nevertheless, SCLs used for the production of biological products such as growth
factors and vaccines should comply with these recommendations.
Some of the general recommendations given here (see sections A.1 to
A.5) are applicable to all animal cell substrates. More specific guidance for PCCs
can be found in the relevant documents published by WHO (e.g. production of
poliomyelitis vaccine in primary monkey kidney cells) (4, 5).
Cell substrates should be developed and used in accordance with
applicable requirements of the NRA/NCL.
In general, it is not consistent with good manufacturing practices (GMP)
to retest materials that have already been released for further manufacture, so
justification would be necessary before such retesting is undertaken. Thus, the
scope of this document is intended to cover cell substrates as new cell banks
are established. The specific circumstances under which the retesting of already
established and released cell banks would be appropriate should be discussed
with the responsible NRA/NCL.
Recommendations published by WHO are intended to be scientific and
advisory in nature. The parts of each section printed in normal type have
been written in the form of recommendations so that, should an NRA/
NCL so desire, this text may be adopted as it stands as the basis of national
or regional requirements. The parts of each section printed in small type
are comments or additional points that may be considered in some cases.
4. Definitions
The definitions given below apply to the terms used in these recommendations.
The terms may have different meanings in other contexts.
Adventitious agent: contaminating microorganisms of the cell culture
or source materials including bacteria, fungi, mycoplasmas/spiroplasmas,
mycobacteria, Rickettsia, protozoa, parasites, transmissible spongiform
89
encephalopathy (TSE) agents, and viruses that have been unintentionally

introduced into the manufacturing process of a biological product.
The source of these contaminants may be the legacy of the cell line, the
raw materials used in the culture medium to propagate the cells (in
banking, in production, or in their legacy), the environment, personnel,
equipment or elsewhere.
Biological medicinal product: biological medicinal product is a synonym

for biological product or biological described in the WHO Technical Report
Series. The definition of a medicinal substance used in treatment, prevention
or diagnosis as a “biological” has been variously based on criteria related to its
source, its amenability to characterization by physicochemical means alone, the
requirement for biological assays, or arbitrary systems of classification applied
by regulatory authorities. For the purposes of WHO, including the current
document, the list of substances considered to be biologicals is derived from
their earlier definition as “substances which cannot be fully characterized by
physicochemical means alone, and which therefore require the use of some
form of bioassay” (16). However, developments in the utility and applicability
of physicochemical analytical methods, improved control of biological and
biotechnology-based production methods, and an increased applicability of
chemical synthesis to larger molecules have made it effectively impossible to base
a definition of a biological on any single criterion related to methods of analysis,
source or method of production. Nevertheless, many biologicals are produced
using in vitro culture systems.
Developers of such medicinal products that do not fit the definition
of biological medicinal product provided in this document should
consult the relevant NRAs for product classification and the licensing
application pathway.
Biotherapeutic: for the purpose of this document, a biotherapeutic is a

biological medicinal product with the indication of treating human diseases.
Cell bank: a collection of appropriate containers whose contents are of
uniform composition, stored under defined conditions. Each container represents
an aliquot of a single pool of cells.
The individual containers (e.g. ampoules, vials) should be representative
of the pool of cells from which they are taken and should be frozen on the
same day by following the same procedure and by using the same equipment
and reagents.
Cell culture: the process by which cells are grown in vitro under defined
and controlled conditions, where the cells are no longer organized into tissues.
90
Annex 3
Cell line: type of cell population with defined characteristics that originates
by serial subculture of a primary cell population that can be banked.
Cloning and subcloning steps may be used to generate a cell line. The
term “cell line” implies that cultures from it consist of lineages of some of
the cells originally present in the primary culture.
Cell seed: a quantity of well-characterized cells that are frozen and stored
under defined conditions, such as in the vapour or liquid phase of liquid nitrogen,
in aliquots of uniform composition derived from a single tissue or cell, one or
more of which would be used for the production of a master cell bank. Cell seed
is also referred to as a pre-MCB or seed stock. It may be made under conditions
of GMP or under the manufacturer’s research and development conditions.
Cell substrate: cells used to manufacture a biological product.
The cells may be primary or cell lines, and may be grown in monolayer
or suspension culture conditions. Examples of cell substrates include
primary monkey kidney, MRC-5, CHO, and Vero cells.
Cells used to generate essential components of a final product (e.g. Vero

cells for the generation of “reverse genetics” virus for use in seeding
vaccine production) are considered to be “preproduction” cell substrates.
Cells used to manufacture the bulk product (e.g. packaging cell lines for
gene therapy vectors, Vero cells for vaccine production, or CHO cells for
recombinant protein expression) are considered to be “production” cell
substrates.
Continuous cell line (CCL): a cell line with an apparently unlimited

capacity for population doubling. It is often referred to as “immortal” and was in
the past referred to as “established”.
Diploid cell line (DCL): a cell line with a finite in vitro lifespan in which
the chromosomes are paired (euploid) and are structurally identical to those of
the species from which they were derived.
While this definition is accurate for standard chromosome preparations,
a given human diploid cell line may contain genetic variations that will
be reflected in a Giemsa-banding pattern that differs from the standard.
Gene expression differences may also be found.
This definition is based on experience and the current understanding of

the in vitro behaviour of human cells that are not of stem cell origin.
DNA infectivity: the capacity of cellular DNA to generate an infectious

virus following introduction of that DNA into appropriate cells. The viral genome
may be integrated or extrachromosomal.
91
Endogenous virus: a virus whose genome is present in an integrated

form in a cell substrate. Endogenous viruses are present in the genome of the
original animal from which the cells were derived. They may or may not encode
an intact or infectious virus.
End-of-production cells (EOPCs): cells harvested at or beyond the end
of a production (EOP) run.
In some cases, production cells are expanded under pilot-plant scale or
commercial-scale conditions.
Extended cell bank (ECB): cells cultured from the MCB or WCB and
propagated to the proposed in vitro cell age used for production or beyond.
Functional integrity: the culture sustains the expected performance
related to its intended use under specified conditions (e.g. expression of secreted
product at a consistent level, production of expected yield of virus).
Immortalized: having an apparently unlimited capacity for population
doubling.
Indicator cells: cells of various species used in the in vitro adventitious
agent test that are intended to amplify adventitious viruses to promote their
detection. Generally, this would include a human diploid cell line (such as
MRC-5), a monkey kidney cell line (such as Vero cells) and a cell line of the same
species and tissue as the cell bank. The purpose of these cell lines is to “indicate”
a viral infection of the cell bank either through observation of cytopathic effect
(CPE) during and after an appropriate observation period or by haemadsorption
and/or haemagglutination at the end of the observation period. Thus they are
referred to as “indicator” cells. The cell bank may be analysed on such indicator
cells either by co-cultivation or by passage of cell lysates or spent culture
supernatants from the cell bank on to the indicator cells.
In vitro cell age: the measure of time between the thaw of the MCB vial(s)
to the harvest of the production vessel, measured by elapsed chronological time,
by population doubling level of the cells, or by passage level of the cells when
subcultivated by a defined procedure for dilution of the culture.
Latent virus: a virus is considered to be latent when the viral genome is
present in the cell without evidence of active replication but with the potential to
be reactivated.
Master cell bank: a quantity of well-characterized cells of animal or other
origin, derived from a cell seed at a specific PDL or passage level, dispensed into
multiple containers, cryopreserved and stored frozen under defined conditions,
such as the vapour or liquid phase of liquid nitrogen in aliquots of uniform
composition. The MCB is prepared from a single homogeneously mixed pool
of cells. In some cases, such as genetically engineered cells, the MCB may be
prepared from a selected cell clone established under defined conditions.
92
Annex 3
Frequently, however, the MCB is not clonal. It is considered best practice for the
MCB to be used to derive working cell banks.
Oncogenicity: the capacity of an acellular agent – such as a chemical,
virus, viral nucleic acid, viral gene(s) or subcellular element(s) – to cause normal
cells of an animal to form tumours.
Oncogenicity is distinct from tumorigenicity (see “Tumorigencity”). The
tumours that arise in an oncogenicity test are of host origin, whereas in
a tumorigenicity test, the tumours are derived from the inoculated cells.
Parental cells: cells that are manipulated to give rise to a cell substrate. For
hybridomas, it is usual also to describe the parental cells as the cells to be fused.
Manipulation may be as simple as the expansion of a primary cell culture

to provide early-passage cells, or it may be a more complex activity such
as developing a hybridoma or transfected clone. Both processes would
provide a cell seed. The parental cells may refer to any stage prior to
the preparation of the cell seed. Examples of a parental cell are WI-38
and MRC-5 at very early passage, Vero at passage 121, and CHO before
the introduction of a DNA construct to produce a recombinant cell. In
certain situations (e.g. myeloma cells), there may be a lineage of identified
stable parental clones; thus, the term “parental cell” would normally refer
to the cells used immediately prior to generation of the cell seed.
Passage: the process of transferring of cells, with or without dilution,

from one culture vessel to another in order to propagate them, and which is
repeated to provide sufficient cells for the production process.
This term is synonymous with “subculture”. Cultures of the same cell
line with the same number of passages in different laboratories are not
necessarily equivalent because of differences in cell culture media, split
ratios, and other variables that may affect the cells. This is a more important
consideration for SCLs and CCLs than for DCLs. Population doubling is
the preferred method of estimating cell-line age and, whenever possible,
should be used instead of passage. However, it also may be appropriate to
quantify culture duration of CCLs by the number of subcultivations at a
defined seeding density at each passage or time in days.
Population doubling: a twofold increase in cell number.

Population doubling level: the total number of population doublings of
a cell line or strain since its initiation in vitro. A formula to use for the calculation
of population doublings in a single passage is:
number of population doublings = log10 (N/No ) × 3.33
93
where N = number of cells in the growth vessel at the end of a period of growth
and No = number of cells plated in the growth vessel (17). It is best to use the
number of viable cells or number of attached cells for this determination.
Primary culture: a culture started from cells, tissues or organs taken
directly from one or more organisms. A primary culture may be regarded as such
until it is successfully subcultured for the first time. It then becomes a cell line if
it can continue to be subcultured at least several times.
Production cell cultures: A collection of cell cultures used for biological
production that have been prepared together from one or more containers from
the WCB or, in the case of PCCs, from the tissues of one or more animals.
Residual cellular DNA (rcDNA): cell substrate DNA present in the
final product.
Specific-pathogen-free (SPF): animals known to be free of specific
pathogenic microorganisms and reared in an environment that maintains that
state. SPF animals are usually raised in biosecure facilities and their health status
is monitored on an ongoing basis. The SPF status simply provides an assurance
that the stock is not infected with the specified pathogens. SPF animals are not
disease free, nor are they disease resistant; they may carry pathogens other than
those from which they are specified to be free.
Stem cell line: a continuous cell line generated from stem cells rather
than from normal or diseased differentiated tissue.
Transmissible spongiform encephalopathy: the transmissible spongiform
encephalopathies (TSEs) are a group of fatal neurodegenerative diseases which
include classical and variant Creutzfeldt–Jakob disease (CJD), Gerstmann–
Sträussler–Scheinker syndrome (GSS), fatal familial insomnia (FFI) and Kuru
in humans, bovine spongiform encephalopathy (BSE) in cattle, chronic wasting
disease (CWD) in mule, deer and elk, and scrapie in sheep and goats.
Tumorigenicity: the capacity of a cell population inoculated into an
animal model to produce a tumour by proliferation at the site of inoculation and/
or at a distant site by metastasis.

Tumorigenicity is distinct from oncogenicity (see “Oncogenicity”).
WHO reference cell bank (RCB): a cryopreserved stock of cells prepared

from a single homogeneous pool of cells prepared under defined conditions and
subjected to characterization tests. The purpose of such a bank is to serve as a
well-characterized cell seed for the preparation of MCBs that will be extensively
characterized by manufacturers and that have a high probability of meeting
these recommendations.
Working cell bank: a quantity of well-characterized cells of animal or
other origin, derived from the MCB at a specific PDL or passage level, dispensed
into multiple containers, cryopreserved and stored frozen under defined
94
Annex 3
conditions, such as in the vapour or liquid phase of liquid nitrogen, in aliquots

of uniform composition. The WCB is prepared from a single homogeneously
mixed pool of cells. One or more of the WCB containers is used for each
production culture.
5. General considerations
5.1 Types of animal cell substrates
5.1.1 Primary cell cultures (PCCs)
PCCs have played a prominent role in the development of biology, and of virology
in particular. Cultures of PCCs from different sources have been in worldwide
use for the production of live and inactivated viral vaccines for human use for
many decades. For example, PCCs of monkey kidney cells have been used for the
production of inactivated and oral poliomyelitis vaccines since the 1950s.
Major successes in the control of viral diseases, such as poliomyelitis,
measles, mumps and rubella, were made possible through the wide use of vaccines
prepared in PCCs, including those from chicken embryos and the kidneys of
monkeys, dogs, rabbits and hamsters, as well as other tissues.
PCCs are viable cells of disaggregated tissues that are initiated as in vitro
cell cultures, usually as adherent cells. Many cell types are present and a primary
culture can be a complex mixture of cells that may be influenced by the process
and conditions under which they were harvested, disaggregated and introduced
to in vitro culture. Not all cells in a primary culture will have the capacity to
replicate. Particular care should be given to establishing highly reproducible
procedures for tissue disaggregation, cell processing and culture initiation, as
well as reproducible culture conditions and nutrition.
PCCs obtained from wild animals usually show a high frequency of viral
contamination. For instance, monkey kidney cell cultures may be contaminated
with one or more adventitious agents, including simian viruses.
If PCCs are necessary for the production of a given biological, the
frequency of contaminated cell cultures can be significantly reduced by screening
the source animals carefully for the absence of such viruses. Viruses can be
detected by molecular tests such as polymerase chain reaction (PCR), and by
looking for the presence of circulating antibodies to those viruses in the source
animals. The use of animals bred in a carefully controlled colony, especially those
that are SPF, is strongly recommended. Nevertheless, as suitable alternative cell
substrates become available, PCCs are less likely to be used in the future. WHO
has promoted the replacement of animals for experimental purposes, both for
ethical reasons (18) and in the interests of progressive improvement in product
safety and quality.
95
5.1.1.1 Advantages of PCCs
■ PCCs are comparatively easy to prepare using simple media and

bovine serum.
■ They generally possess a broad sensitivity to a variety of viruses,
some of which are cytopathic.
5.1.1.2 Disadvantages of PCCs
■ Contamination by infectious agents is a higher risk than with DCLs

and CCLs.
■ The quality and viral sensitivity of cultures obtained from different
animals are variable.
■ Although cell cultures derived from non-human primates have been
widely used in the past, it has become increasingly difficult to obtain
and justify the use of such animals for this purpose.
■ PCCs cannot be tested as extensively as DCLs or CCLs.
5.1.2 Diploid cell lines (DCLs)

The practicality of using human DCLs for the production of viral vaccines was
demonstrated in the 1960s. Experience gained with oral poliomyelitis and other
viral vaccines in successfully immunizing billions of children in many countries
has shown clearly that such substrates can be used in the production of safe and
effective vaccines (3).
The essential features of DCLs of human (e.g. WI-38, MRC-5) and
monkey (i.e. FRhL-2) origin are:
■ they are cells passaged from primary cultures that have become
established as cell lines with apparently stable characteristics over

numerous PDLs;
■ they have a finite capacity for serial propagation which ends in
senescence, a state in which the culture ceases to replicate; the cells
remain alive and metabolically active but may show morphological
and biochemical changes, some of which begin to appear before
replication ceases;
■ they are non-tumorigenic;
■ they display diploid cytogenetic characteristics with a low frequency
of chromosomal abnormalities of number and structure until they
enter senescence.
96
Annex 3
Substantial experience has been accumulated on the cytogenetics of

WI-38 and MRC-5 since the 1960s, and ranges of expected frequencies of
chromosomal abnormalities have been published (19, 20). Similar data are
available for FRhL-2 (21). More sophisticated cytogenetic techniques (e.g.
high-resolution banding, comparative genome hybridization) (22, 23) have
demonstrated the presence of subtle chromosomal abnormalities that were
previously undetectable. Recent studies have shown that subpopulations of
human DCLs with such abnormalities may appear and disappear over time,
that they are non-tumorigenic and that they undergo senescence in the same
manner as the dominant population. Thus, possessing a stable karyotype might
not be such an important characteristic as was previously thought.
5.1.2.1 Advantages of DCLs
■ DCLs can be well characterized and standardized.

■ Production can be based on a cryopreserved cell bank system that
allows for consistency and reproducibility of the reconstituted cell
populations.
■ A cell bank system usually consists of cell banks of defined population
doubling or passage levels that generally include an MCB and a WCB.
■ Unlike the CCLs and SCLs discussed below, DCLs are not tumorigenic
and therefore do not raise the potential safety issues associated with
CCLs and SCLs.
5.1.2.2 Disadvantages of DCLs
■ DCLs are not easy to use in large-scale production, although they

have been cultivated using bioreactor technology employing the
microcarrier or multilayer method.
■ In general, they have more fastidious nutritional requirements than
other cell substrates.
■ They may be difficult to adapt to serum-free growth.
■ DCLs are more difficult to transfect and engineer than CCLs, and
DCLs require immortalization before they can be engineered (e.g.
they are not permissive for the production of vaccine vectors that
require complementation, since they cannot be engineered readily
to express complementing proteins).
5.1.3 Continuous cell lines (CCLs)

Some CCLs have been used for the production of safe and effective biotherapeutics
and vaccines since the 1980s.
97
CCLs have the potential for an apparently indefinite in vitro lifespan and
have been derived by the following methods:
■ serial subcultivation of a PCC of a human or animal tumour (e.g.
HeLa cells);
■ transformation of a normal cell having a finite lifespan, with an
oncogenic virus or viral sequence (e.g. B lymphocytes transformed
by EBV or transfected with viral sequences such as in PER.C6);
■ serial subcultivation of a primary cell population derived from
normal tissue that generates a dominant cell population having
an apparently indefinite lifespan, often described as spontaneous
transformation (e.g. Vero, BHK-21, CHO, MDCK, Hi5);
■ fusion between a myeloma cell and an antibody-producing B
lymphocyte to produce a hybridoma cell line;
■ use of ectopically expressed genes involved in the cell cycle, such as
hTERT telomerase gene, to enable indefinite replication of normal
human cells.
CCLs may display a consistent modal chromosome number (e.g. MDCK,
Vero), and although the karyotype of individual cells in a culture at one point
in time may vary, the range of chromosome numbers per cell will usually show
characteristic limits. However, other CCLs, such as highly tumorigenic cells
including HeLa, may show variation in modal number and a wider drift in the
range of the number of chromosomes per cell.
In the early stage of establishing a cell line, significant diverse karyotypes
and changes in karyotype may be observed. However, a characteristic chromosome
component may emerge with continued passage, presumably as a dominant cell
population develops.
5.1.3.1 Advantages of CCLs

■ CCLs can be characterized extensively and their culture conditions

can be standardized.
■ Production of CCLs can be based on a cell bank system, which
allows for consistency and reproducibility of the reconstituted cell
populations for an indefinite period.
■ CCLs generally grow more easily than DCLs, using standard media.
■ Most CCLs can be adapted to grow in a serum-free medium.
■ CCLs can usually be grown on microcarriers for large-scale
production in bioreactors.
■ Some can be adapted to grow in suspension cultures for large-scale
production in bioreactors.
98
Annex 3
5.1.3.2 Disadvantages of CCLs
■ CCLs may express endogenous viruses, and some are tumorigenic in

immunosuppressed animal models.
■ Theoretical risks identified by the 1986 Study Group (e.g. nucleic acids,
transforming proteins and viruses) need to be taken into account.
5.1.4 Stem cell lines (SCLs)

Stem cells differ from other types of cells because they sustain a predominant
stem cell population while simultaneously retaining the capacity to produce cell
progenitors of differentiated cell types of almost all human tissues (i.e. they are
pluripotent). Pluripotent SCLs have an apparent capacity to generate cell types of
all three human germ layers and may be capable of generating in vitro models of
any tissue in the human body. At the time these Recommendations were written
(2010), two types of pluripotent SCLs – human embryonic stem cells and induced
pluripotency stem cell lines – had been isolated and were considered to possibly
have the capabilities to prove useful for manufacturing biologicals. The property
of pluripotency is sustained through numerous cycles of cell division. SCLs may
be derived from early-stage embryonic, fetal or adult tissues. Typically, specialized
media and environmental conditions such as the attachment matrix are required
for the growth of SCLs in vitro, in order for them to maintain the undifferentiated
state. While most stem cell research and development has been directed towards
transplantation of stem cells for therapeutic purposes, efforts also have been made
to explore a variety of SCLs as cell substrates for the production of biologicals.
Key considerations for the culture and control of such cell lines have been
developed (24). These include the fundamental issues common to the maintenance
of all cell lines, but also include the need for appropriate ethical governance
regarding donor consent and careful attention to periodic confirmation of
phenotype, absence of non-diploid cells, and sustained pluripotent capacity.
It has recently been shown that conditioned medium from SCLs can have
regenerative properties. Such preparations produced from human embryonic
stem cells have shown regenerative capabilities, including repair of myocardial
infarction in animal models (25). This raises the possibility of stem cells being
used as a substrate to produce a variety of biologically active molecules. SCLs can,
in some respects, be considered as diploid cells, but they do not appear to have
the finite lifespan characteristic of human diploid fibroblast cultures. In human
embryonic stem cell cultures, clonal variants with chromosomal abnormalities
are known to arise. While a diploid and non-transformed nature is considered a
prerequisite for cell therapy applications, transformed SCLs might be considered
as a form of CCL for the manufacture of biologicals. Because they do not fall
easily within any one category of substrate already discussed, SCLs are identified
separately in this document.
99
5.1.4.1 Advantages of SCLs
■ SCLs can be well characterized and their culture conditions

standardized.
■ Production can be based on a cell bank system, allowing consistency
and reproducibility of the reconstituted cell populations for an
indefinite period.
■ Some SCLs may be adapted to grow in suspension cultures for large-
scale production in bioreactors.
■ SCLs may produce unique proteins of potential importance as
biotherapeutics.
■ SCLs have the potential to generate cells and tissue-like structures
that may permit the expression of agents that are currently
considered unculturable in vitro.
5.1.4.2 Disadvantages of SCLs
■ Subculture techniques commonly used for SCLs are laborious.

■ SCLs may produce growth proteins with undefined effects on adult
cells/tissues.
■ SCLs usually require complex media that may have a TSE risk.
■ Rapid development of differentiated cells also means that they are
difficult to control in vitro.
■ There is little experience with the use of SCLs as a cell substrate to
manufacture biological products.
5.2 Potential risks and risk mitigation associated with

biologicals produced in animal cell cultures
The main potential risks associated with the use of biologicals produced in animal
cells are directly related to contaminants from the cells. These risks fall into three
categories, namely:
■ viruses and other transmissible agents;
■ cellular nucleic acids (DNA and RNA);
■ growth-promoting proteins.
In addition, cell-derived inhibiting or toxic substances are theoretically
possible. A summary of the risk assessment for each follows. More comprehensive
information has been published elsewhere on the risks associated with
contaminating DNA and growth-promoting proteins (26–35).
100
Annex 3
In 2010, NRAs and WHO were made aware of new information regarding
the presence of DNA sequences of porcine circovirus in live-attenuated rotavirus
vaccines. The detection of these sequences by the use of advanced analytical
methods raised complex questions (e.g. the evaluation of the potential risk, specific
testing of vaccines, and the general use of these methods for the characterization
of vaccine cell substrates). The power of the new methodology that was used
(i.e. massively parallel (deep) sequencing (MPS)) may reveal the presence of
adventitious agents that might not be detected with current methods. While the
implementation for routine use of such methods has benefits as well as challenges
and risks, NRAs need to be prepared for similar situations. Consideration should
be given to making a risk assessment and potentially introducing risk-mitigation
strategies in such circumstances.
5.2.1 Viruses and other transmissible agents

There is a long history of concern regarding the potential transmission of viruses
and other infectious agents that may be present in cell substrates. WHO reviewed
this area in 1986 through a Study Group that pointed out that, as described
below, cells differ with respect to their potential for carrying viral agents that are
pathogenic for humans.
Primary monkey kidney cells have been used to produce billions of
doses of poliomyelitis vaccines since they were first developed in the 1950s and,
although viruses such as SV40 were discovered in rhesus monkey kidney cells,
control measures were introduced to eliminate or reduce as much as possible the
risk of viral contamination associated with the manufacture of vaccines in cells
containing those viruses. Additional controls may be needed as new viral agents
are identified and technologies to detect them are developed.
Human and non-human primate lymphocytes and macrophages may
carry latent viruses such as herpesviruses and retroviruses. CCLs of non-
haematogenous cells from human and non-human primates may contain viruses
or have viral genes integrated into their DNA. In either case, virus expression
may occur under conditions of in vitro culture.
Avian tissues and cells may harbour exogenous and endogenous
retroviruses, but there is no evidence of transmission of disease to humans from
products prepared using these substrates. For example, large quantities of yellow
fever vaccines were produced for many years in eggs that contain avian leukosis
viruses, but there is no evidence that these products have transmitted disease in
their long history of use for human immunization. Nevertheless, the potential for
transmitting avian retroviruses should be reduced as much as possible, through
control measures during manufacture.
Rodents may harbour exogenous and endogenous retroviruses,
lymphocytic choriomeningitis virus, and hantaviruses, and a range of other
101
potentially zoonotic viruses. While contamination with the rodent viruses in

the cell harvests of biotherapeutic products derived from CHO cell culture has
been reported (35–37), there is no evidence that biological products released for
distribution have been contaminated with rodent viruses. If present, such viruses
were detected during quality-control testing in compliance with GMP prior to
release. In addition, it is important to note that there have been no reported
cases of transmission of an infectious agent to recipients of recombinant protein
products manufactured in animal cells.
Insect cells have recently been used for vaccine production, and various
insect cell lines may be used for the production of biologicals in the future. Insect
viruses tend to be ubiquitous in many insect cell lines and are generally unknown
and/or uncharacterized. Many insect cell lines have endogenous transposons
and retrovirus-like particles, and some are positive in product-enhanced reverse
transcriptase (PERT) assays.
HDCs have been used for vaccine production for many years. Although
concern was initially expressed about the possibility of such cells containing a
latent pathogenic human virus, no evidence for such an endogenous agent has
been reported, and vaccines produced from this class of cell substrate have
proven to be free from viral contaminants.
In light of the differing potential of the various types of cells mentioned
above for transmitting viruses that are pathogenic in humans, it is essential
that the cells being used to produce biological products should be evaluated as
thoroughly as possible with respect to infectious agents.
When DCLs, SCLs or CCLs are used for production, a cell bank system
should be used and the cell banks should be characterized as specified in this
document. Efforts to identify viruses by testing for viral sequences or other viral
markers, especially those not detectable by other means, constitute an important
part of the evaluation of cell banks in addition to the standard tests that have
been in place for many years.
When cell lines of rodent or avian origin are examined for the presence of
viruses, the major emphasis in risk assessment should be on the results of studies
in which transmission to target cells or animals is attempted. Risk to human
recipients should not be assessed solely on ultrastructural or biochemical/
biophysical evidence of the presence of viral or viral-like agents in the cells.
The overall manufacturing process – including the selection and testing
of cells and source materials, any purification procedures used, and tests
on intermediate or final products – should ensure the absence of detectable
infectious agents in the final product. When appropriate, validation of purification
procedures should demonstrate adequate reduction of relevant model viruses,
with a significant safety factor (14). This is usually required for recombinant
protein products.
102
Annex 3
There may be as yet undiscovered microbial agents for which there is

no current evidence or means of detection. As such agents are identified, it will
be important to consider whether to re-examine cell banks for their presence.
In general, it is not a practice consistent with GMP to retest materials that have
already been released, so justification would be necessary before such retesting is
undertaken. Positive findings should be discussed with the NRA/NCL. Whenever
new data are developed with the potential for impact on the quality, safety or
efficacy of a biological product, it is the responsibility of the manufacturer to
provide NRAs with all the relevant data and information that are currently
available. This should include confirmation and evaluation of the finding, the
manufacturer’s own risk assessment and an investigational and action plan, in
order to facilitate any regulatory action that may be necessary. In addition, new
testing methods are likely to be developed and, as they become available and
validated, they should be considered by manufacturers and NRAs/NCLs for their
applicability to the characterization and control of new animal cell substrates.
5.2.2 Cellular DNA

The issue of rcDNA in biological products has been considered by many
groups since the 1980s, and there has been an evolution of consensus on
recommendations. The most recent WHO Recommendation (WHO Technical
Report Series, No. 878) (1) sets the upper limit of rcDNA at 10 ng per parenteral
dose. As stated below, while this value has proved helpful in the past, it does not
take into consideration important factors such as the size of the DNA fragments
and any potentially inactivating steps in the manufacturing process. Thus, it is
important to take into consideration not only the limit of 10 ng per parenteral
dose but other factors as well, when determining the acceptable level of rcDNA.
PCCs and DCLs have been used successfully for many years for the
production of viral vaccines, and the rcDNA deriving from these cells has not
been (and is not) considered to pose any significant risk. However, with the use
of CCLs, which have an apparently indefinite lifespan, presumably due to the
dysregulation of genes that control growth, and with the ongoing development
of products from cells that are tumorigenic or were derived from tumours, the
DNA from such cells has been considered to have the theoretical potential to
confer the capacity for unregulated cell growth, and perhaps oncogenic activity,
upon some cells of a recipient of the biological product. Although the risk of
such DNA has been estimated on the basis of certain assumptions and some
experimental data, assessing the actual risk of such DNA has not been possible
until recently, when preliminary data generated from new experimental systems
began to quantify the risk (38).
The potential risk of DNA arises from both of its biological activities,
namely infectivity and oncogenicity. Infectivity could be due to the presence
103
of an infectious viral genome in the cellular DNA of the cell substrate (39–
41). The viral genome could be that of a DNA virus, whether integrated or
extrachromosomal, or of a proviral genome of a retrovirus. Both types of viral
DNA have been shown to be infectious in vitro and, in several cases, in vivo (39,
40). The oncogenic activity of DNA could arise through its capacity to induce a
normal cell to become transformed and perhaps to become tumorigenic. The
major mechanism through which this could occur would be the introduction
of an active dominant oncogene (e.g. myc, activated ras), since such dominant
oncogenes could directly transform a normal cell. Other mechanisms would
require that the rcDNA transforms through insertional mutagenesis, and
have been considered less likely since the frequency of integration of DNA is
generally low (42). The frequency of integration at an appropriate site, such as
inactivating a tumour suppressor gene or activating a proto-oncogene, would be
correspondingly lower (32).
The 1986 WHO Study Group addressed the risk posed by the oncogenic
activity of rcDNA in biological products for human use (12). Risk assessment
based on a viral oncogene in an animal model suggested that in vivo exposure
to 1 ng of rcDNA, where 100 copies of an activated oncogene were present in the
genome, would give rise to a transformational event once in 10 9 recipients (27).
On the basis of this and other evidence available at that time, the Study Group
concluded in 1986 that the risk associated with rcDNA in a product is negligible
when the amount of such DNA is 100 pg or less per parenteral dose. On the basis
of a review of more recent data, those requirements were revised in 1998 to raise
the acceptable level of rcDNA to 10 ng per dose.
Studies in mice using cloned cellular oncogenes also suggest that the
risk of neoplastic transformation by cellular DNA is probably very low (34, 43).
However, more recent data have shown that cloned cellular oncogene DNA can
induce tumours in selected strains of mice at levels below 1 ng. In addition,
single oncogenes can also be biologically active (44) and can initiate the tumour
induction process. Because of these data and the recent evidence that genes
encoding for certain micro-RNA species can be oncogenic in vitro (45–48), thus
increasing the number of potential dominant cellular oncogenes, the oncogenic
risk of DNA needs to be taken into account when tumorigenic cells are considered
for use in the production of biologicals. This would be especially important for
live attenuated viral vaccines where chemical inactivation of the DNA is not
possible and where the only way to reduce the biological activity of DNA would
be by nuclease digestion and the reduction in the quantity of DNA.
In addition to its oncogenic activity, the infectivity of DNA should be
considered. Since a viral genome, once introduced, could amplify and produce
many infectious particles, the infectivity risk is likely to be greater than the
104
Annex 3
oncogenic risk. The polyoma virus genome is infectious in mice at about 50 pg

(49), and a recent report demonstrated that 1 pg of a proviral copy of a retrovirus
is infectious in vitro (50). Because such low levels of DNA may be biologically
active, the amounts of rcDNA should be factored into safety evaluations when
tumorigenic cell substrates are used, especially for live viral vaccines.
Consequently, considerations that need to be taken into account with
respect to rcDNA are: (i) any reduction in the amount of the contaminating DNA
during the manufacturing process; (ii) any size reduction of the contaminating
DNA during the manufacturing process; and (iii) any chemical inactivation of
the biological activity of contaminating DNA during the manufacturing process.
A product might be considered by an NRA/NCL to have an acceptable level of
risk associated with the DNA of the cell substrate on the basis of (i) and/or (ii)
and/or (iii) when data demonstrate that appropriate levels have been achieved.
For example, data have shown that nuclease digestion of DNA or chemical
inactivation of DNA with beta-propiolactone, a virus-inactivating agent, can
destroy the biological activity of DNA (38, 50, 51). Therefore, the use of these
procedures may provide an additional level of confidence with respect to
reduction of DNA risk.
For products such as monoclonal antibodies and subunit vaccines
manufactured in tumorigenic cell substrates, it is necessary to demonstrate the
clearance (removal and/or inactivation) of DNA by the manufacturing process.
This may require validation of the main inactivating or removal steps. For
example, data should be obtained on the effects of DNA-inactivating agents
under specific manufacturing conditions, so that firm conclusions can be drawn
on their DNA-inactivating potential for a given product.
There may be instances where CCL DNA is considered to pose a higher
level of risk because it contains specific elements such as infectious retroviral
proviral sequences. Under these circumstances, the steps taken to reduce the
risks of rcDNA, such as reducing the size of DNA fragments, should be agreed in
consultation with the NRA/NCL.
The 1986 WHO Study Group stated that the risks for rcDNA should be
considered to be negligible for preparations given orally. This conclusion was based
on the finding that polyoma virus DNA was not infectious when administered
orally (49). For such products, the principal requirement is the elimination of
potentially contaminating viruses. Recently, additional data have been published
on the uptake of DNA via the oral route (52). These studies demonstrated that
the efficiency of uptake of DNA introduced orally was significantly lower than for
that introduced intramuscularly. Nevertheless, the specifics of the manufacturing
process and the formulation of a given product should be considered by the
NRA/NCL.
105
With respect to the efficiency of DNA uptake via the nasal route,
no data have been published comparing this route with parenteral routes.
However, data suggest that uptake via the intranasal route is less efficient than
by the intramuscular route (53). Limits for a specific product should be set in
consultation with the NRA/NCL.
In general, acceptable limits of rcDNA for specific products should
be agreed in consultation with the NRA/NCL, taking into consideration the
characteristics of the cell substrate, the intended use of the biological product and,
most importantly, the effect of the manufacturing process on the size, quantity
and biological activity of rcDNA fragments. In general, it has been possible to
reduce rcDNA in biotechnological products to <10 ng per dose, and in many
cases to <10 pg per dose, because they can be highly purified. Quantitative PCR
for short amplicons has been used to determine total residual DNA levels as well
as residual DNA fragment size distribution. It should be noted that other methods
may give different results for small fragments or for DNA that has been treated
with inactivating agents. Whatever methods are used should be validated. Some
products, especially certain live viral vaccines, are difficult to purify without
a significant loss in potency, so the amount of rcDNA in those final products
may be significantly higher than 10 ng per dose. Such cases are considered to be
exceptional and should be discussed with the NRA/NCL.
5.2.3 Cellular RNA

While protein-coding RNA has not been considered to be a risk factor for
biological products, owing to the unstable nature of RNA and the lack of
mechanisms for self-replication, the recent description of small non-coding RNA
molecules – microRNA (miRNA) – that are more stable and have the capacity to
modulate gene expression may necessitate a reassessment. Whether these miRNA
molecules can be taken up by cells in vivo is unknown. However, as previously
stated, because certain miRNA genes can be oncogenic, DNA containing such
sequences may need to be considered along with oncogenes, when assessing the
risk of rcDNA (see section B.9 on Oncogenicity). Because this is an evolving area
of research, no conclusions can be made regarding the risk of miRNA and no
recommendations are made to control miRNA at this time.
5.2.4 Growth-promoting proteins

Growth factors may be secreted by cells that are used to produce biologicals, but
the risks from these substances are limited since their growth-promoting effects
are usually transient and reversible, they do not replicate, and many of them are
rapidly inactivated in vivo. In exceptional circumstances, growth factors may
contribute to oncogenesis, but even in these cases, the tumours apparently remain
106
Annex 3
dependent upon continued administration of the growth factor. Therefore, the

presence of known growth factor contaminants at ordinary concentrations
does not constitute a significant risk in the preparation of biological products
manufactured in animal cell cultures. However, some SCLs may secrete higher
levels and more potent factors than CCLs. This should be taken into account
when designing characterization studies. The manufacturing process should be
designed to address any safety issues that are identified.
Part A. General recommendations applicable to

all types of cell culture production
A.1 Good manufacturing practices
The general principles of GMP for biologicals should be in place. Requirements
or recommendations have been made by NRAs (e.g. the European Medicines
Agency (EMA), the United States Food and Drug Administration (FDA)) and
other groups (e.g. the International Conference on Harmonisation of Technical
Requirements for Registration of Pharmaceuticals for Human Use (ICH)). GMP
should be applied from the stage of cell banking onwards.
In the preparation of a cell substrate, it is considered best practice to
establish tiered master and working cell banks (see section A.5.2), to ensure a
reliable and consistent supply of cells that can be fully characterized and safety
tested prior to use for production. By definition, PCCs cannot be subjected to
such banking regimens. However, some manufacturers have utilized pooled and
cryopreserved primary cultures, which enable completion of lot release testing as
in a tiered banking system. The strategy for delivery of primary cells or primary
cells recovered from cryopreservation should be based on the quality and safety
that can be assured for the final product according to the overall manufacturing
and control processes involved.
A.2 Principles of good cell culture practice

A.2.1 Understanding the cells and the culture system
In all aspects of sourcing, banking and preparing cell cultures, the principles of
good cell culture practice should be observed (e.g. (54, 55) for SCLs). Fundamental
features to be considered in the development of cell cultures for production or
testing are:
■ authenticity, including identity, provenance and genotypic/phenotypic

characteristics;
■ absence of contamination with another cell line;
107
■ absence of microbiological contamination;

■ stability and functional integrity on extended in vitro passage.
An important basic principle for all types of cells is that the donor should
be free of transmissible diseases or diseases of uncertain etiology, such as CJD
for humans and BSE for cattle. The NRA/NCL may allow specific exceptions
concerning donor health (e.g. myeloma, other tumour cells).
Cells in culture may change their characteristics in response to changes
in culture conditions or on extended passage under the same culture conditions.
The four cell culture types (PCC, DCL, SCL, CCL) used in manufacture differ
in their potential stability. Thus, characterization approaches may need to be
adapted to reflect these differences.
Cell cultures grow in an in vitro environment that is substantially different
from the conditions experienced by cells in vivo, and it is not unexpected that
they may be susceptible to change or alteration as a result of in vitro culture and
processing. It is important to be conscious of the variation that may arise in the
cell culture environment, as cells may undergo subtle alterations in their cell
biology in response to such changes. It is therefore necessary to try to control
key known variables that could have significant impact on cell culture. Medium
and specific additives (serum, growth factors, amino acids and other growth-
promoting compounds) should, where possible, be specified in terms of chemical
composition and purity. Where relevant, the biological activity of the medium
and the additives should be determined before use. New batches of reagents for
cell culture should be supplied with certificates of analysis and origin, to enable
their suitability to be evaluated against the established specification. The use of
serum or other poorly defined reagents is not recommended in the production
of new biologicals from cell culture, and chemically defined alternatives should
be sought wherever possible. However, given that our current understanding
of cell biology is not complete, the benefits that defined media bring in the
form of higher reproducibility and reduced risk must be balanced against the
potential effects of inadequacies of defined culture systems that may not meet
the full biological needs of cells. Where complex biological reagents such as
fetal bovine serum (FBS) remain necessary, they should be carefully controlled
whenever possible by pre-use selection of batches. Such careful selection also
should apply, where relevant, to cell culture surfaces using specified culture
vessels or surface coatings.
Variation in physical culture parameters – such as pH, temperature,
humidity and gas composition – can significantly influence the performance
and viability of cells and should be specified with established tolerances, and the
relevant equipment calibrated and monitored. In addition, any culture reagents
prepared in the laboratory should be documented, controlled for quality and
released against an established specification.
108
Annex 3
A.2.2 Manipulation of cell cultures

In vitro processing of cells can introduce additional physical and biochemical
stresses that could have an influence on the quality of the final product. Care
should be taken to minimize manipulations, taking into consideration the
specifics of the manufacturing process. In all cases, a consistent process should
be demonstrated.
A.2.2.1 Detachment and subculture

Detachment solutions may adversely affect the cells if exposure is not minimized.
Cell harvesting and passaging procedures should be carried out in a reproducible
way that ensures consistency in the confluency of cells when harvested, in
incubation times, temperature, centrifugation speeds and times, and in post-
passage viable cell seeding densities.
A.2.2.2 Cryopreservation
(see section A.5.1)
A.2.2.3 Introduction of contamination

The microbiological status of the donor individual, colony, herd or flock is
an important consideration in the establishment of PCCs. In order to avoid
catastrophic failure of the production process and to avoid infectious hazards
for the recipients of products, it is important to minimize the opportunities
for contamination of cell cultures. Therefore, cell manipulation and open
processing steps should be minimized, taking into consideration the specifics of
the manufacturing process. It is critical to adopt a rigorous aseptic technique
and to provide appropriate environmental controls and air quality for cell
culture processing and the preparation of growth media. The presence of any
antimicrobial in a biological process or product is discouraged, although a
notable exception is that antibiotic(s) and antifungal(s) may be required for
primary cell cultures. Additionally, antibiotics may be used in some cell line
selection systems. Where antibiotics have been used, sterility-testing procedures
should take into account the potential inhibitory effects of the antibiotic on
contaminating organisms. Penicillin or other beta-lactam antibiotics should not
be present in production cell cultures.
A.2.3 Training and staff

Training in all cell culture processes is vital to ensure that correct procedures are
adhered to under GMP. Staff should be trained in the underlying principles of cell
culture procedures, to give them an understanding of cell culture processes that
will enable them to identify events and changes that could influence the quality of
109
cells (54). Key procedures on which training in good cell culture practice should
focus include the passaging of cells, preparation of sterile media, and maintenance
and use of biological safety cabinets, incubators and cryopreservation.
Cell cultures should be prepared by staff who have not, on the same
working day, handled animals or infectious microorganisms. Furthermore, cell
cultures should not be prepared by staff who are known to be suffering from a
transmissible infection. The personnel concerned should undergo a “return to
work” assessment to evaluate any residual risk.
A.2.4 Cell line development and cloning

Wherever a cell culture has passed through a process that may have a significant
influence on its characteristics, such as tumorigenicity, it should be treated
as a new (i.e. different) cell line and should be renamed with a suffix or code
to identify this. An MCB should then be prepared from the “post-treatment”
culture. Treatments that may require such rebanking include cell cloning and
genetic manipulation. Any change(s) to the cell culture process should be
demonstrated not to affect product quality and should be discussed with the
NRA. In the manufacture of monoclonal antibodies, cloning of hybridoma
cultures is particularly important in order to ensure that a single product is
generated, since inclusion of more than one hybridoma cell type could lead to a
mixture of different antibody specificities and classes being present.
The details of cloning and selection may vary according to the practices
of individual manufacturers and should be discussed with the NRA/NCL.
In the early stages of cell line development, a number of different
recombinant vector systems and cell lines may be used. This will essentially be
a research activity, but the cell lines and vectors should originate from well-
characterized and qualified sources and the cells from an appropriately qualified
seed stock or MCB, which will usually be “in-house” host cells and vectors.
The most promising cell–vector combination will then be used to generate a

large number of clones (from a few hundred to thousands) after transfecting
the culture with rDNA. Typically, these clones will be screened on the basis of
their productivity, and a number with the highest productivity (10–50) will be
taken forward for further evaluation. Further testing will then be used to select a
small number (1–5) for establishment as small pre-master cell banks, and a final
selection will be made – often based on stability characteristics and amenability
to scale-up – before finally an MCB and WCB are generated. Throughout the
process, only well-characterized and traceable growth media and other critical
reagents will be used (usually the same as for the MCB), and cryopreserved
stocks of all working clones will be made at appropriate stages in the development
process (see Figure A3.1).
110
Annex 3
Figure A3.1
Simplified outline of the development of a genetically modified cell line
In the process of cloning a cell culture, single cells should be selected for
expansion. The cloning procedure should be carefully documented, including the
provenance of the original culture, the cloning protocol and reagents used. Cloning
by one round of limiting dilution will not necessarily guarantee derivation from
single cells; additional subcloning steps should be performed. Alternatively, or in
addition to limiting dilution steps, the cloning procedure can include more recent
technology such as single-cell sorting and arraying or colony-picking from dilute
seeds into semi-solid media. In any case, the cloning procedure should be fully
documented, with details of imaging techniques and/or appropriate statistics. For
proteins derived from transfection with recombinant plasmid DNA technology,
a single fully documented round of cloning is sufficient, provided that product
homogeneity and consistent characteristics are demonstrated throughout the
production process and within a defined cell age beyond the production process.
It is important to document accurately the establishment of each clone,
which should also have a unique reference. Cryopreserved seed stocks of a
significant number of clones should be established at an early stage. The clones
can then be compared in parallel with the parental culture, to establish candidate
clones with the best overall characteristics for delivery of the desired product.
The criteria used in the evaluation of the clone selected for production should
include genomic and phenotypic stability, growth rate, achievable product levels,
and integrity/stability of the product. The evaluation of early candidate clones
should generate sufficient information for the manufacturer to make an informed
decision on the selection of the most promising clone(s) for further development.
111
Where genetically engineered cell clones are under evaluation, these criteria
should also include the stability of integrated rDNA. The details of this process
could vary due to a number of factors, including the nature of the host cell, the
desired characteristics of the product, and the manufacturer’s local procedures.
It is important to bear in mind that, even following single-cell cloning,
epigenetic variation may result in a cloned culture showing evidence of
heterogeneity (i.e. more than one clone). This should not preclude the use of
such a culture for production unless there are indications of instability that could
affect the quality and/or safety of the final product.
A.2.5 Special considerations for neural cell types

Agents causing TSEs have been propagated in certain cells. At the time of writing,
the phenomenon has been observed only with very specific pairs of agents
and cells; no cell line has been identified that will replicate all agents, although
one has been described that seems to be infectable by many strains of scrapie.
Although the phenomenon is unpredictable, if the line does not express the prion
protein (PrP), it may be assumed to be impossible to infect it, and experience
to date suggests that infection is not commonly observed or easy to maintain.
On the other hand, the cell types that can be infected include fibroblastic lines
as well as neuronal cells. The cells have usually been of murine origin, because
the infecting agents are usually mouse-adapted scrapie. The fact that certain
cells can be infected with certain agents is proof of the principle that cell lines
may be infected; thus, exposure of cells to sources potentially contaminated
with the agents is a concern. Since the scale of the risk is difficult to judge, it
is recommended that, with respect to safety considerations and TSEs, attention
should focus on the selection and documentation of the cell culture reagents and
other materials that come into intimate contact with the cells, in order to provide
assurance that they are not contaminated. Strategies to accomplish this are given in
section B.11.4.
A.3 Selection of source materials

A.3.1 Introduction
All materials should be subjected to risk assessment and testing when necessary
– particularly raw materials derived from humans and animals, which can be a
primary source for the introduction of adventitious agents into the production of
biologicals. Careful attention should be paid to sourcing, production, handling,
testing and quality control. All cell culture materials of biological origin that come
into intimate contact with the cells – during the establishment of cell cultures,
derivation of a new cell line (if any), banking procedures (if any) and production –
should be subject to appropriate tests, as indicated by risk assessment, to establish
112
Annex 3
quality and freedom from contamination by microbial agents and to evaluate

acceptability for use in production. It is important to evaluate the microbiological
risks represented by each human- and animal-derived reagent used in a cell
culture production process. The evaluation should address: (i) geographical
origin; (ii) the species of origin; (iii) general microbiological potential hazards,
including a consideration of the medical history for human-derived reagents;
(iv) the husbandry/screening of donor animals; (v) testing performed on the
product, including certificates of analysis (if any); and (vi) the capacity for the
preparation, purification and sterilization procedures (if any) used to remove
or inactivate contaminants (56). Other reagents of biological but non-animal
origin may also present risks to product safety, and these are discussed further
in section A.3.4.
Recombinant protein technology now provides many materials that were
formerly derived directly from animal or human sources. While this eliminates
obvious virological risks from donors, the manufacturing process used for the
recombinant proteins should be analysed to identify any materials of biological
origin and any associated hazards that may need to be addressed.
The NRA/NCL should approve the source(s) of animal-derived raw
materials such as serum and trypsin. These materials should comply with the
guidelines on tissue infectivity distribution of TSEs (57). The materials should
be subjected to appropriate tests for quality and freedom from contamination
by microbial agents, to evaluate their acceptability for use in production. Their
origin should be documented to ensure that the sources are from geographical
regions with acceptable levels of microbiological risk (e.g. freedom from
foot-and-mouth disease virus or BSE). In addition, documentation should be
gathered on their manufacturing history, production, quality control and any
final or supplementary processing that could affect quality and safety (such
as blending and aliquoting of serum batches). Controls should be in place to
prevent cross-contamination of one material with another (e.g. bovine material
in a porcine product).
The reduction and elimination from the manufacturing process of raw
materials derived from animals and humans is encouraged, where feasible.
For some human- and animal-derived raw materials used in the cell
culture medium, such as insulin or transferrin, validation of the production
process for the elimination of viruses can substitute for virus detection tests,
when justified.
Animal-derived reagents such as trypsin and serum, which would be
substantially damaged or destroyed in physical sterilization processes, including
heat and irradiation, present the most likely microbiological hazards to cell
culture processes. Batches of reagents, such as trypsin and bovine serum, have
been known to contain Mycoplasma species and sometimes more than one
viral contaminant. Certain contaminants have also been shown to infect cells
113
in culture. The processing environment is a common source of microbiological

contamination and should be controlled to minimize this risk and to prevent
growth of contaminants.
A.3.2 Serum and other bovine-derived materials used in cell culture media
The source(s) of serum of bovine origin should be approved by the NRA/NCL.
The responsibility for ensuring the quality of the serum used in the manufacture
of cell banks and biologicals rests with the manufacturer of the biologicals.
Quality can be ensured in more than one manner. The manufacturer may
conduct testing for adventitious agents and perform inactivation of the serum
after purchase from the serum manufacturer. Alternatively, the manufacturer
may qualify the serum vendor and purchase serum from suppliers only after
conducting thorough and ongoing audits of the serum suppliers to ensure that
they have properly performed the manufacture, quality control and validation
necessary to achieve the level of serum quality required for the biological being
produced. In some cases, certificates of analysis may then be considered sufficient.
Some combination of these approaches might be optimal, and the strategy taken
should be considered when evaluating risk. Consultation with the NRA/NCL
may also be advisable.
Serum and other bovine-derived materials should be tested for
adventitious agents such as bacteria, fungi, mycoplasmas and viruses, prior
to use in the production of MCBs and WCBs and in the manufacture of
biologicals. Particular consideration should be given to those viruses that could
be introduced from bovine-derived materials and that could be zoonotic or
oncogenic (e.g. bovine viral diarrhoea virus (BVDV), bovine polyoma virus,
bovine circoviruses, rabies virus, bovine adenoviruses, bovine parvovirus (BPV),
bovine respiratory syncytial virus (BRSV), infectious bovine rhinotracheitis virus
(IBR), bovine parainfluenza virus type 3 (BPIV3), reovirus 3 (REO3), Cache
Valley virus, bluetongue virus (BTV) and epizootic haemorrhagic disease
virus). Consideration should also be given to risk-mitigation strategies, such as

inactivation by heat or irradiation, to ensure that adventitious agents that are not
detected in the manufacture and quality control of the serum will be inactivated
to a degree that is acceptable to the NRA/NCL. If irradiation or other inactivation
methods (e.g. heat sterilization) are used in the manufacture of the serum, the tests
for adventitious agents should be performed prior to inactivation, to enhance the
opportunity for detecting contamination. If evidence of viral contamination is
found in any of the tests, the serum is acceptable only if the virus is identified and
shown to be present in an amount that has been shown in a validation study to
be effectively inactivated. If evidence of viral contamination is found in any tests
on serum that is not to be subjected to a virus inactivation or removal procedure,
the serum would not generally be acceptable. If the manufacturer chooses to
use serum that has not been inactivated, thorough testing of the serum for
114
Annex 3
adventitious agents, using current best practices, should be undertaken. If any

agents are identified, the cell banks made in this manner must be shown to be free
of the identified virus(es).
If irradiation is used, it is important to ensure that a reproducible dose
is delivered to all batches and to the component units of each batch.
The irradiation dose must be low enough for the biological properties
of the reagents to be retained, while being high enough to reduce
virological risk. Thus, irradiation delivered at such a dose may not be a
sterilizing dose.
If serum was used in the establishment or passage history of the animal

cell substrate prior to banking by the manufacturer, the cell bank (MCB or WCB)
and/or cells at or beyond the level of production should be tested for adventitious
agents of the species (e.g. bovine) of serum used in the establishment and
passage history of the cell substrate. If serum is not used in the production of
the subsequent stages, then this testing would not need to be repeated on those
subsequent stages, once the cell bank has been tested and considered free of
bovine (or whichever species of serum was used) adventitious agents.
Methods used to test for bovine viruses should be approved by the
NRA/NCL. Details of the methods are provided in Appendix 1. Infectivity
assays are used as the primary screening method and have resulted in the
detection of BVDV, REO3, Cache Valley virus, BTV and epizootic haemorrhagic
disease virus, among others. However, it should be noted that, in general, the
infectivity screening assay methods described here do not readily detect some
of the viruses (e.g. bovine polyomaviruses) that can be frequent contaminants of
serum. Additional methods may need to be considered, such as the nucleic acid
amplification technique (NAT), although the presence of viral genomic sequences
is not necessarily indicative of infectious virus. In those cases, specific infectivity
assays designed to detect the virus of concern (e.g. bovine polyomavirus) may
need to be considered.
A second factor in screening serum is the limited sample volume used,
compared with the batch size (which may be of the order of 1000:1) that comes
from the pooling of serum from many animals. Consequently, infectious
viruses may be missed in the serum lot testing, and consideration should be
given to direct screening of the cell bank for bovine viruses. These assays could
include, in addition to the general screening procedure, NAT for the presence
of bovine viruses that may infect the cell substrate but undergo abortive and/or
transforming infections. Virus families of particular concern in this regard include
polyomaviruses, herpesviruses, circoviruses, anelloviruses and adenoviruses.
General screening assays for the detection of infectious viruses in serum
or cell substrates involve the use of at least one indicator cell line, such as bovine
turbinate cells, that is permissive for the replication of BVDV. A second cell line,
115
such as Vero, should also be employed to broaden the detection range. Before
initiating screening, it may be necessary to evaluate the serum for the presence of
antibody, particularly to BVDV, that could mask the presence of infectious virus.
Indicator cells are cultured typically in the presence of the serum under
test for 21–28 days, passaging the cells as necessary. During this period, the cells
are regularly examined for the presence of CPE indicative of virus infection. At
the end of the observation period, which should not be less than 7 days after the
last subculture, the cells are stained to detect CPE that may have been missed
during observation of the living cells. Additional end-point assays should include
haemadsorbtion and haemagglutination at both 4 °C and a higher temperature
such as 20–25 °C and also immunofluorescence assays (IFAs) for specific
viruses. Appropriate controls should be used for each assay – such as BPIV3 for
haemadsorbtion. IFAs are particularly important for BVDV, since non-cytopathic
BVDV may be present in the serum. IFA end-points are also used to detect
other viruses that may be determined by geographical considerations – such as
adenoviruses, BPV, BTV, BRSV, REO3 and unlikely but serious contaminants
like rabies virus.
If serum from another species is used (i.e. other than bovine), the NRA/
NCL should be consulted about acceptable testing methods for that species.
A.3.3 Trypsin and other porcine-derived materials

used for preparing cell cultures
Trypsin used for preparing cell cultures should be tested for cultivable bacteria,
fungi, mycoplasmas and infectious viruses, including bovine or porcine
parvoviruses, as appropriate. The methods used to do this should be approved
by the NRA/NCL.
In some countries, irradiation is used to inactivate potential contaminant
viruses. If irradiation is used, it is important to ensure that a reproducible
dose is delivered to all batches and the component units of each batch.
The irradiation dose must be low enough for the biological properties of
the reagents to be retained, while being high enough to reduce virological
risk. Thus, irradiation cannot be considered a sterilizing process.
The quality of the trypsin, like serum, is the responsibility of the

biologicals manufacturer (see section A.4.2). Recombinant trypsin is available
and should be considered. However, it should not be assumed that recombinant
trypsin is free from the risk of contamination and it should be subject to the usual
considerations for any reagent of biological origin (see section A.4.1).
Like serum batches, which are derived from many animals, trypsin
batches are also prepared from the pancreases of many animals. Most batches
of porcine trypsin contain genetic sequences of porcine parvovirus 1 and
porcine circoviruses and should therefore be treated in a manner accepted by the
116
Annex 3
NRA/NCL in order to inactivate any virus that may potentially be present. It is

acknowledged, however, that these viruses are relatively resistant to inactivation
(58). If trypsin from another species is used, the NRA/NCL should be consulted
regarding acceptable testing methods.
General screening assays for the detection of infectious porcine viruses in
trypsin or cell substrates involve the use of at least one indicator cell line, such as
porcine testes cells or Vero cells, that is permissive for the replication of porcine
viruses. Typically, indicator cell cultures would be incubated for 14 days with
a subculture on to fresh test cells for an additional 14 days. Specific end-point
detection methods such as IFA or PCR may be required, in addition to periodic
observation for CPE throughout the culture period and more general end-point
detection methods such as haemadsorption and/or haemagglutination.
If trypsin has been used in the establishment or passage history of the
animal cell substrate prior to banking by the manufacturer, the cell bank (MCB
or WCB) should be tested for porcine parvovirus or for appropriate adventitious
viruses relevant to the species of origin of the trypsin used. If trypsin is not used
in the production of the subsequent stages, then this testing would not need to
be repeated on those subsequent stages once the cell bank has been tested and is
considered free of porcine parvovirus (or relevant agents). Consideration should
be given to screening for other agents such as porcine circoviruses. Molecular
methods such as PCR may be used for such purposes.
Testing of cells exposed to trypsin or of other porcine-derived materials
might entail testing for more than porcine parvovirus or porcine circoviruses.
For instance, it may be appropriate to test for porcine adenovirus, transmissible
gastroenteritis virus, porcine haemagglutinating encephalitis virus, BVDV,
reoviruses, rabies virus, porcine anellovirus, porcine hokovirus, porcine bocavirus,
porcine hepatitis E virus, porcine reproductive and respiratory syndrome virus,
encephalomyocarditis virus and potentially other viruses. Particular consideration
should be given to those viruses that could be introduced from the porcine-
derived material and that could be zoonotic or oncogenic. Additionally, tests for
bacterial and fungal sterility and mycoplasmas should be conducted depending
on the type of porcine-derived material. The NRA/NCL should be consulted on
this issue.
A.3.4 Medium supplements and general cell culture reagents derived

from other sources used for preparing cell cultures
The quality of medium supplements derived from other species should be
controlled from the perspective of adventitious agents. Consideration should
be given to whether recombinant-derived medium supplements were exposed
to animal-derived materials during their manufacture and, if so, they should be
evaluated for the potential to introduce adventitious agents into the manufacture
of the cell banks and biological products. Testing for adventitious agents should
117
assess viruses relevant to the species from which the supplement was derived.
The NRA/NCL should be consulted on this issue.
Medium supplements should generally not be obtained from human
source materials. In particular, human serum should not be used. However, in
special circumstances, and in agreement with the NRA/NCL, the use of human-
derived supplements may be permitted. If human serum albumin is used at any
stage of product manufacture, the NRA/NCL should be consulted regarding
the requirements, as these may differ from country to country. As a minimum,
human serum albumin should meet the revised Requirements for biological
substances no. 27 (59), as well as the Guidelines on tissue infectivity distribution
of TSEs (57).
Recombinant human albumin is commercially available and should be
considered. However, it should not be assumed to be free of risk of contamination
and should be subject to the usual safety considerations for any reagent of
biological origin (see section A.3.1).
As for other cell culture reagents, it is important to establish traceability
and to assess and reduce microbiological risks as described in section A.3.1.
A variety of cell culture reagents of biological origin is available. These
reagents are derived from non-animal sources, including a range of aquatic
organisms, plants and algae. In such cases, the exact hazards involved may be
uncertain and unfamiliar. The microbiological risks may be substantially different
from those involved in animal-derived reagents (see sections A.3.1–A.3.3),
and other hazards may arise – such as immunogenic, mitogenic and allergenic
properties of the reagent and its components. Plant-derived material may, for
instance, carry an increased risk of mycoplasma and mycobacteria contamination.
A.4 Certification of cell banks by the manufacturer

It is vital that the manufacturer secures a body of information on the cell substrate,
demonstrating clearly the origin or provenance of the culture and how the cell
banks intended for production (MCB and WCB) were established, characterized
and tested. This should provide all the information required to demonstrate the
suitability of that cell substrate and the established cell banks for the manufacture
of biological products.
A.4.1 Cell line data

Each new cell line should have an associated body of data, which will increase
as the cell line is established and developed for manufacturing purposes. This
dataset is vital for demonstrating the suitability for use of the cells and should
provide information on cell provenance (donor information and any relevant
details on ethical procurement), cell line derivation, culture history, culture
conditions (including reagents), early-stage safety evaluation data, banking and
118
Annex 3
cell bank characterization, and safety testing. This information should be made
available to the NRA/NCL for approval of the cells used in manufacture.
A.4.2 Certification by the manufacturer of primary cell cultures

Full traceability of PCCs should be established to the animals of origin,
husbandry conditions, veterinary inspection, vaccinations (if any), procedures
for administering anaesthesia and cell harvesting, reagents and procedures used
in the preparation of primary cultures, and the environmental conditions under
which they were prepared. Extensive testing should be performed and should
be documented.
It is important to define the batch or lot of cells used in each individual
manufacturing process. For production purposes, a batch or lot is a culture of
primary cells derived from single or multiple animals that has been subjected
to a common process of tissue retrieval, disaggregation and processing, leading
to a single-culture preparation of cells. Lots may be prepared by harvesting and
pooling cells in different ways, but the cell-processing procedures should be
reproducible. It is especially important to monitor cultures carefully for evidence
of adverse change in the cell culture and of microbiological contamination.
Prior to any culture pooling, cells should be examined for acceptability for
production. Acceptability criteria should be established and should include
testing for microbiological contamination and the general condition of the cells
(e.g. morphology, number, viability of the cells). Failure to detect and eliminate
atypical (i.e. potentially virally infected) or grossly contaminated cells will put the
entire production run at risk and could compromise the safety of the product.
Cells showing an unacceptably high proportion of dead or atypical cells should
not be used, and microbiological testing should ideally be completed and passed
before the cells are used.
The preparation of cell lots for manufacture should be carefully
documented, to provide full traceability from the animal donor(s) to production.
Any pooling of cells should be clearly recorded, as should any deviation from
standard operating procedures, as required by GMP. In addition, any observations
of variation between batches should be recorded, even where such observations
would not necessarily lead to rejection of those batches. Such information may
prove valuable in ongoing optimization and improvement of the production
process.
A.4.3 Certification by the manufacturer of diploid,

continuous and stem cell lines
All cell lines used for production of biologicals should have data available, as
indicated in section A.4.1. The original PDL (or passage number, if the PDL is
unknown) of the cell seed should be recorded. For cell lines of human origin, the
medical history of the individual from whom the cell line was derived should, if
119
possible, be evaluated in order to better assess potential risks and the suitability
of the cell line.
For SCLs, morphology continues to be an important characteristic.
Representative images and immune-phenotypic profiles of undifferentiated and
differentiated cells should be available for comparison.
A.5 Cryopreservation and cell banking

A.5.1 Cryopreservation
When cells are banked, the successful preservation of cells at ultra-low temperatures
is critical to the efficient delivery of good quality cultures (i.e. high-viability
cultures with the required characteristics). The need to prepare large stocks of
frozen vials of cells for cell banks is especially challenging, and a number of key
principles should be adopted.
■ A method that meets current best practice for cell culture
preservation should be used (see, for example (60)).
■ The cooling profile for the cells being frozen should be defined,
and the same cooling process should be used for each separate
preservation process (i.e. the standard operating procedure should
include documentation of the cooling process in the batch record).
■ Each preservation process should be recorded.
■ As a general guide, only cell cultures that are predominantly in the
exponential phase of growth should be used. Cells in such cultures
tend to have a low ratio of cytoplasm to nucleus (v/v) and should be
more amenable to successful cryopreservation. It is unwise to use
cells predominantly in the “lag” phase very early after passage, or
in the “stationary” phase when the culture has reached its highest
density of cells.
■ For each bank, cells pooled from a single expanded culture (i.e. not
from a range of cultures established at different times post-seeding

or different PDLs) should be used and mixed prior to aliquoting to
ensure homogeneity.
■ The number of cells per vial should be adequate to recover a
representative culture (e.g. 5–10 × 10 6 in a 1 ml aliquot).
■ For new cell banks, antimicrobials should not be used in cell cultures
to be banked, except where this can be justified for early PDL
cultures, which may carry contamination from tissue harvesting
or recombinant cells which require antibiotic selection, and when
necessary for the genetic stability of recombinant cell lines. If
antimicrobials are used, they should not be penicillin or any other
beta-lactam drug.
120
Annex 3
■ When a stock of cells has been frozen, a sample should be recovered

to confirm that it has retained viability and the results have been
recorded. It is also important to establish the degree of homogeneity
within the cell bank. Recovery of a sufficient percentage (e.g. 1%,
or as recommended by the NRA) of vials representative of the
beginning, middle and end of the cryopreservation process should
be demonstrated to give confidence in the production process based
on the use of that cell bank. Ultimately, stability (see section B.3)
and integrity of cryopreserved vials is demonstrated when the vials
are thawed from production and shown to produce the intended
product at scale (see also section B.7).
■ Cell bank cryostorage vessels should be monitored and maintained to
enable demonstration of a highly stable storage environment for cell
banks. Access to such vessels may cause temperature cycling, which
in extreme cases can cause loss of viability. It is prudent to establish
a stability-testing programme that involves periodic recovery of cells
where the frequency of recovery relates to the risk of temperature
cycling. New developments in remote monitoring of individual vials
may, in the future, eliminate the need for stability testing.
A.5.2 Cell banking

When DCLs, SCLs or CCLs are used for the production of a biological, a cell
bank system should be in place. Cell banks should be approved by and registered
with the NRA/NCL as part of the product approval process. The source of cells
used in cell banking and production is a critical factor in biological product
development and manufacture. It is highly desirable to obtain cells from sources
with a documented history and with traceability to the originator of the cell line.
After a sample of the original seed stock is obtained, an early-stage pre-
master bank of just a few vials should be established. One or more vials of the pre-
master bank are used to establish the MCB. The WCB is derived by expansion of
one or more containers of the MCB. The WCB should be qualified for yielding
cell cultures that are acceptable for use in manufacturing a biological product.
When using early PDLs from primary cultures for production processes,
the preparation of a cell bank should be considered on a case-by-case basis. This
approach has significant benefits: it gives great flexibility in the timing of the
production process, permits quality control and safety testing to be completed
prior to use, and reduces the overall burden of testing required in the process.
Cell banks should be characterized as specified in Part B of these
recommendations, and according to any other currently valid and future guidance
published by WHO. The testing performed on a replacement MCB (derived from
the same cell clone or from an existing MCB or WCB) is the same as for the initial
MCB, unless a justified exception is made. Efforts to detect contaminating viruses
121
and other microbial agents constitute a key element in the characterization of

cell banks.
Having been cryopreserved by qualified methods (see section A.5.1),
both the MCB and WCB should be stored frozen under defined conditions,
such as in the vapour or liquid phase of liquid nitrogen. The location, identity
and inventory of individual cryovials or ampoules of cells should be thoroughly
documented. It is recommended that the MCB and WCB each be stored in at least
two widely separated areas within the production facility and/or in geographically
distinct locations, in order to ensure continued ability to manufacture products
in the event of a facility catastrophe. When cryopreserved cells are transferred to
a remote site, it is important to use qualified shipping containers and to monitor
transfers with probes to detect temperature excursions. All containers are treated
identically and, once removed from storage, are not usually returned to the stock.
The second storage site should operate under an equivalent standard of quality
assurance to that at the primary site.
A.5.3 WHO reference cell banks (RCBs)

The principle of establishing RCBs under WHO auspices is one that offers
potential solutions to future challenges for the development of vaccines and
biotherapeutics in developing regions. However, WHO does not intend that
cells supplied to manufacturers from any RCB should be used as an MCB. The
purpose of WHO RCBs is to provide well-characterized cell seed material for the
generation of an MCB by manufacturers, with the expectation that such MCBs
will comply with this guidance document and be fully characterized.
The WHO RCBs provide key advantages for vaccine development
worldwide, which include:
■ traceability to the origin of cells and derivation of the cell line and
materials used in preparation of seed stock;
■ being subject to open international scientific scrutiny and collaborative

technical investigations into the characteristics of the cells and the
presence of adventitious agents;
■ the results of characterization are peer-reviewed and published;
■ investigations are evaluated under the auspices of WHO expert review
and qualified as suitable for use in vaccine production;
■ supply of cells free of any constraint related to intellectual property
rights on final products;
■ a single source of cells with a growing and scientifically and technically
updated body of safety-testing data and safe history of use, giving
increased confidence to manufacturers, regulators and public
policy-makers.
122
Annex 3
The Vero cell line has been the most widely used continuous cell line
for the production of viral vaccines over the past two decades. The WHO Vero
RCB 10-87 was established in 1987 and was subjected to a broad range of tests
to establish its suitability for vaccine production. This WHO RCB provides
a unique resource for the development of future biological medicines where a
cell substrate with a history of safe and reliable use is desired. A comprehensive
review of the characterization of the WHO Vero 10-87 seed lot was conducted
recently, and a detailed overview is provided on the WHO web site (http://www.
who.int/biologicals/).
As concluded by an expert review in 2002, the WHO Vero RCB 10-87 is
not considered suitable for direct use as MCB material. However, the WHO Vero
RCB 10-87 is considered suitable for use as a cell seed for generating an MCB,
and its status has changed from “WHO Vero cell bank 10-87” to “WHO Vero
reference cell bank 10-87”.
The WHO Vero RCB 10-87 is stored in the European Collection of Animal
Cell Cultures (www.hpacultures.org.uk) in the United Kingdom and the American
Type Culture Collection (ATCC, www.atcc.org) in the USA. These public service
culture collections have distributed ampoules under agreements with WHO to
numerous manufacturers and other users. The WHO Vero RCB 10-87 is the
property of WHO and there are no constraints relating to intellectual property
rights. The WHO Vero RCB 10-87 is available free of charge on application to
WHO. However, owing to the limited number of vials remaining, distribution
is restricted to use in the production of vaccines and other biologicals. Potential
replacement of the WHO Vero RCB 10-87 is currently under consideration.
WHO also has overseen the establishment of seed stocks of MRC-5 for the
production of vaccines. The WHO MRC-5 RCB was established in 2007 because
of stability issues associated with the original vials of MRC-5 cells, which dated
from 1966. This RCB was prepared in a qualified cleanroom environment and
was subjected to specified quality-control testing endorsed by the WHO Expert
Committee on Biological Standardization.
Part B. Recommendations for the characterization

of cell banks of animal cell substrates
B.1 General considerations
Since the 1986 Study Group report, advances in science and technology have
led to an expanded range of animal cell types being used for the production
of biologicals. In some cases, these new cell types provide significantly higher
yields of product at less cost, while in other cases they provide the only means
by which a commercially viable product can be manufactured. Many products
123
manufactured in CCLs of various types have been approved, and some examples
are listed in Table A3.1.
Table A3.1
Examples of approved biological products derived from CCLs
Product class Product (disease) Cell line

Therapeutic Factor VIII (haemophilia) CHO
Factor VIIa (haemophilia) BHK-21
Monoclonal antibody (various CHO and murine myeloma
diseases) (NS0 and SP2/0)
Prophylactic Poliovirus vaccine Vero
Rotavirus vaccine Vero
Rabies vaccine Vero
Japanese encephalitis vaccine Vero
Human papillomavirus vaccine Sf-9
Influenza vaccine MDCK
Many more products are currently in development. Some products use

highly tumorigenic cells (e.g. HeLa, and some banks of MDCK) and some involve
sources previously unused in production, such as insect cells. Examples are listed
in Table A3.2.
Table A3.2
Examples of biological products in development derived from CCLs
Product class Product (disease) Cell line(s)

Therapeutic More than 50% of products in development use CHO or murine
myeloma cells as the cell substrate (61). Monoclonal antibodies are
generally produced using CHO, SP2/0, PER.C6 and NS0 cells (62).
Prophylactic HIV vaccines CHO, Vero, PER.C6, 293ORF6,
HER96, HeLa
Herpes simplex type 2 vaccine CHO
Influenza vaccines sf9, Vero, PER.C6
Rabies vaccine S2
124
Annex 3
CCLs may have biochemical, biological and genetic characteristics that

differ from PCCs or DCLs and that may impose a risk for the recipients of the
biologicals derived from them. In particular, CCLs may produce transforming
proteins and may contain potentially oncogenic DNA and viral genes. In some
cases, CCLs may cause tumours when inoculated into animals. Non-tumorigenic
cells (e.g. PCCs and DCLs) had been thought to be intrinsically safer than
tumorigenic cells. Where tumorigenic cells have been used in the past (e.g. CHO
for recombinant proteins), high degrees of purity have been required, with a
special emphasis on reduction in the quantity of DNA. When it was not possible
to reduce the amount of DNA to below the detection limit, emphasis has been
placed on a reduction in size or other approaches to inactivate rcDNA and rcRNA
(e.g. beta-propiolactone for rabies vaccine).
Manufacturers considering the use of CCLs should be aware of the need
to develop, evaluate and validate efficient methods for purification, as an essential
element of any product development programme. However, a minimally purified
product, such as certain viral vaccines (e.g. polio), may be acceptable if produced
in a CCL such as Vero, when data are developed to support the safety of the
product. Such data would include extensive characterization of the MCB or the
WCB and of the product itself.
While tumorigenicity tests have been part of the characterization of
CCLs, they comprise only one element in an array of tests, the results of which
must be taken into account when assessing the safety of a biological produced
in a given cell substrate. For example, if a CCL is positive in a tumorigenicity
test, and if the CCL is to be used for the production of a live viral vaccine, an
evaluation of the oncogenic potential of the cells may be requested by the NRA/
NCL, to characterize the cellular DNA and to detect oncogenic viruses that might
be present. Such studies should be discussed with the NRA/NCL.
Evidence should be provided for any animal cell line that is proposed for
use as a substrate for the manufacture of a biological product, to demonstrate, to
the limits of the assay’s detection capabilities that the CCL is free from cultivable
bacteria, mycoplasmas, fungi and infectious viruses, including potentially
oncogenic agents. Special attention should be given to viruses that commonly
contaminate the animal species from which the cell line is derived, and to cell
culture reagents of biological origin. The cell seed should preferably be free of all
microbial agents. However, certain CCLs may express endogenous retroviruses.
Tests capable of detecting such agents should be carried out on cells grown
under cell culture conditions that mimic those used during production, and the
levels of viral particles should be quantified. Viral contaminants in an MCB and
WCB should be shown to be inactivated and/or removed by the purification
procedure used in production. The validation of the purification procedure used
is considered essential.
125
The characterization of any DCL, SCL or CCL to be used for the

production of biologicals should include:
■ a history of the cell line (i.e. provenance) and a detailed description of

the production of the cell banks, including methods and reagents used
during culture, PDL, storage conditions, viability after thawing, and
growth characteristics;
■ the results of tests for infectious agents;
■ distinguishing features of the cells, such as biochemical,
immunological, genetic or cytogenetic patterns, that allow them to
be clearly distinguished from other cell lines;
■ the results of tests for tumorigenicity, including data from the
scientific literature.
Additional consideration should be given to products derived from cells

that contain known viral sequences (e.g. Namalwa, HeLa, 293 and PER.C6).
The recommendations that follow are intended as guidance for NRAs,
NCLs and manufacturers, as the minimum amount of data on the cell substrate
that should be available when considering a new biological product for approval.
The amount of data that may be required at various stages of clinical development
of the product should be discussed and agreed with the NRA/NCL at each step
of the programme.
B.2 Identity
Cell banks should be authenticated by a cell identification method approved by
the NRA/NCL. Wherever practicable, methods for identity testing should be
used that give specific identification of the cell line, in order to confirm that no
switching or major cross-contamination of cultures has arisen during cell banking

and production. A number of the commonly used identity-testing methods
are compared in Table A3.3. In the case of human cells, genetic tests such as
DNA profiling (e.g. short tandem repeat analysis, and multiple single nucleotide
polymorphisms) will give a profile that is at least specific to the individual from
whom the cells were isolated. Another test that may be used for human cells
is human leukocyte antigen (HLA) typing. Other tests that may be used but
tend to be less specific include isoenzyme analysis and karyology, which may be
particularly useful where there are characteristic marker chromosomes. However,
where more specific genetic markers are available, they should be considered.
A small proportion of cell lines – particularly those that are transformed –
may show alterations to the expected identity profile. This has been observed
126
Annex 3
in isoenzyme analysis where, in rare cases, a particular cell line may show a
consistently different profile from that expected for the species of origin; it is also
a general issue relating to the effect of genetic instability on molecular identity-
testing techniques. Such effects in standard technologies are rare but may also arise
with the implementation of new techniques. The implications of any unexpected
results should be discussed with the NRA/NCL. For recombinant protein
products, cell-line identity testing should also include tests for vector integrity,
expression plasmid copy number, insertions, deletions, number of integration
sites, percentage of host cells retaining the expression system, verification of
protein-coding sequences, and protein-production levels.
Table A3.3
Identity testing for mammalian cell lines
Technique Advantages Disadvantages

Karyology (especially Gives whole chromosomal Results are generally not
useful for DCLs and SCLs) genome visualization and specific to the individual
analysis that can identify of origin (i.e. there is
species of origin for a very usually species specificity),
wide range of species using although certain cell
the same methodology. lines may have marker
chromosomes that are
readily recognized.
Karyology (especially Newer methods include Giemsa banding requires

useful for DCLs and SCLs) spectral karyotyping, special expertise and is
which involves the use labour intensive.
of probes labelled with Standard analysis of
fluorescent dyes. The probes 10–20 metaphase spreads
paint the chromosomes, is insensitive for detecting
yielding different colours contaminating cells.
in specific areas. Spectral
karyotyping is able to detect
translocations that are not
recognizable by traditional
banding methods.
Isoenzyme analysis Determination of species of Analysis for 4–6

origin within a few hours. isoenzyme activities will
generally identify the
species of origin but is not
specific to the individual
of origin.
127
Table A3.3 continued

Technique Advantages Disadvantages
DNA profiling using Short tandem repeats (STR) Some limited, but
variable number of analysis by PCR is rapid and undefined, cross-reaction
tandem repeats (VNTR) gives identity specific to the of human STR primers for
analysis or other PCR individual of origin. primate cells.
method such as exon-
Commercial kits are
primed intron crossing-
available for a range of
PCR (EPIC-PCR), or
human populations.
techniques such as
restriction fragment The EPIC-PCR method is
length polymorphism rapid and gives identity
(RFLP) specific to the individual
of origin. It provides the
advantage of covering
a broad spectrum of
organisms and cell lines
other than human cells.
Applicability
■ Cell banks: MCB and each WCB

■ Cell types: DCL, SCL, CCL
B.3 Stability
The stability of cell banks during cryostorage, and the genetic stability of cell lines
and recombinant expression systems, are key elements in a successful cell bank
programme.
B.3.1 Stability during cryostorage

Data should be generated to support the stability or suitability of the cell substrate
and any recombinant expression system or necessary cell phenotype during
cultivation to or beyond the limit of production, and to support the stability of
the cryopreserved cell banks during storage. The latter may be demonstrated by
successful manufacture of WCBs or production lots. Periodic testing for viability
is not necessary if continuous monitoring records for storage show no deviations
out of specification, and periodic production runs are successful. If banks are
used less than once every 5 years, it may be prudent to generate data to confirm
their suitability for manufacturing on a schedule that takes into account the
storage condition once every 5 years.
128
Annex 3
Applicability
■ Cell banks: MCB and WCB

B.3.2 Genetic stability

Any form of genetic instability can potentially affect the quality of the final
product and it is important to know if the cells in culture are changing in a way
that could affect the nature or safety of the product. Any features of the cell lines
that might affect quality should be discussed with the NRA/NCL to ensure that
tests used by the manufacturer to monitor genetic stability are adequate. The
actual tests will vary according to the nature of the product, but the aim is to
show consistency in the amount and characteristics of the product derived from
cells within a few passages of the MCB or WCB with those derived from an ECB
or EOPC. For recombinant protein products, emphasis will be on the protein
sequence and post-translational modifications.
For cell lines containing DNA-expression constructs, the stability of these
constructs between the MCB/WCB and an ECB or EOPC should be determined.
The copy number of the construct and, if relevant, the sites of chromosomal
insertion, should also be determined. The latter is accomplished by sequencing into
the cellular flanking regions, but methods such as fluorescent in situ hybridization
may provide useful additional information, particularly where concatamers of
the gene insert are present at individual chromosomal loci. The sequence of the
construct within the cells should be determined. With conventional sequencing,
a consensus sequence is obtained, but with MPS it is possible to determine the
sequence of individual gene inserts or their transcripts.
Where proteins are derived from cells that have not been genetically
modified, consistency in the yield and properties of the protein should be
evaluated, together with the sequence of the messenger RNA (mRNA) encoding
the protein of interest.
Additional characterization of the cell biological processes and responses
during cultivation (for instance, using global or targeted gene expression,
proteomic or metabolic profiles and other phenotypic markers) might be useful
in further developing a broad understanding of the cell substrate.
Appropriate methods should be applied to ensure that cell age is correctly
assessed in the event that cell viability falls dramatically at any given step. Losses
in viability are reflected in increased cultivation times to reach defined levels
of growth.
The stability of cell function in terms of productivity within the
production process also may need to be evaluated. Other stability studies may be
performed where bioreactor methods are employed, especially where extended
culture periods are involved.
129
Applicability
■ Cell banks: MCB taken to EOPC/ECB

B.4 Sterility
(see section B.11.3.1)
B.5 Viability
A high level of viability of cryopreserved cells is important for efficient and
reliable production. Thawed cells should typically have viability levels in excess
of 80%, though this is not always achieved and may depend on the cell line.
Lower viabilities may still result in suitable growth recovery and in acceptable
product qualities. In such cases, the data should be discussed with the NRA/
NCL. A range of viability tests is available to measure different attributes of cell
function (e.g. membrane integrity, metabolic activity, DNA replication). Under
certain circumstances, such as pre-apoptotic cells excluding trypan blue, viability
assays may give misleading results and it is important to be aware of the exact
information that a particular viability assay provides. Therefore, it is important to
evaluate the growth recovery of cryopreserved cells upon thawing.
For certain cell cultures such as hybridomas, where a membrane-integrity
test is used, additional cell markers such as indicators of apoptosis should
be studied, in order to avoid significant overestimation of viability.
A suitable viability test should be selected for the cell substrate in question
and typical test values established for cultures considered to be acceptable (see
sections B.6 and B.7). It may also be necessary to select alternative viability assays
that are better suited to providing the in-process viability data that are required
during production (e.g. lactate dehydrogenase levels in bioreactor systems).
Applicability
B.6 Growth characteristics

For the development of production processes, the growth characteristics of the
production cell line should be well understood, in order to ensure consistency
of production. Changes in these characteristics could indicate any one of a
range of events. Accordingly, data on growth characteristics – such as viability,
morphology, cell-doubling times, cloning and/or plating efficiency – should be
130
Annex 3
developed. For certain cell substrates, it may be appropriate to apply such tests in
homogeneity testing (see section B.7). Experiments to demonstrate homogeneity
and growth characteristics may be combined, although the analysis should be
carried out separately.
Applicability
B.7 Homogeneity
Each cell culture derived from a container of the WCB should perform in the
same way (i.e. within acceptable limits) and should provide the same number
of viable cells with the same growth characteristics. In order to ensure this, it is
important to recover a proportion of containers from each cell bank and check
their characteristics, as indicated in section B.6. The number of containers tested
should be discussed with the NRA/NCL and should be broadly in line with
the number normally sampled to establish product consistency. Recovery of a
sufficient percentage of vials representative of the beginning, middle and end of
the aliquoting process should be demonstrated, in order to give confidence in the
production process that is based on the use of that cell bank. Ultimately, stability
and integrity of cryopreserved vials are demonstrated when the vials are thawed
for production and are shown to produce the intended product at scale. Instead
of testing a proportion of containers at different stages of the banking process,
an alternative strategy to ensure the homogeneity of the banks can be based
on the validation of the process method for filling and freezing. Assessment of
growth characteristics (see section B.6) and homogeneity testing are commonly
combined experimentally; however, the analysis and interpretation of each
should be distinct. It may also be appropriate to test the homogeneity of the MCB
to assure that future WCBs are consistent with the first WCB.
Applicability
■ Cell banks: MCB, WCB

B.8 Tumorigenicity
B.8.1 General considerations
Several in vitro test systems, such as cell growth in soft agar (63) and muscle organ
culture (64), have been explored as alternatives to in vivo tests for tumorigenicity.
131
However, correlations with in vivo tests have been imperfect, or the alternative
tests have been technically difficult to perform. Thus, in vivo tests remain the
standard for assessing tumorigenicity.
Although WHO Requirements (1) have described acceptable approaches
to tumorigenicity testing, a number of important aspects of such testing were
not addressed. Therefore, a model protocol has been developed and is appended
to this document (see Appendix 2). The major points included in the model
protocol are listed below, along with comments on each item.
A new DCL (i.e. other than WI-38, MRC-5 and FRhL-2) should be tested
for tumorigenicity as part of the characterization of the cell line, but should not
be required on a routine basis.
The tumorigenicity tests that are currently available are in mammalian
species, whose body temperatures and other physiological factors are different
from those of avian and insect species. Therefore, when the test is performed
on avian or insect cells, the validity of the data is open to question unless a
tumorigenic cell line of the species being tested is included as a positive control.
The NRA/NCL may accept the results of an in vitro test, such as growth in soft
agar, as a substitute for the in vivo test for avian and insect cell lines. However,
as mentioned above, correlations of in vitro tests with in vivo tests are imperfect.
This should be discussed with the NRA/NCL.
Many CCLs (e.g. BHK-21, CHO, HeLa) are classified as tumorigenic
because they possess the capacity to form tumours in immunosuppressed animals
such as rodents. Some CCLs become tumorigenic at high PDLs (e.g. Vero),
although they do not possess this capacity at the lower PDLs at which vaccine
manufacture occurs. A critical feature regarding the pluripotency of embryonic
SCLs, even though they display a diploid karyotype, is that they form tumours in
immunocompromised mice.
The expression of a tumorigenic phenotype can vary from one CCL
to another, and even within different sublines of the same CCL. This range of
variability, from non-tumorigenic, to weakly tumorigenic, to highly tumorigenic,
has been viewed by some as indicating different degrees of risk when the CCLs
are used as substrates for the manufacture of biological products (10, 11).
If the CCL has already been demonstrated to be tumorigenic (e.g. BHK-21,
CHO, HEK293, Cl27), or if the class of cells to which it belongs is tumorigenic (e.g.
hybridomas, SCLs), it may not be necessary to perform additional tumorigenicity
tests on cells used for the manufacture of therapeutic products. Such cell lines
may be used as cell substrates for the production of biologicals if the NRA/
NCL has determined, on the basis of characterization data and manufacturing
data, that issues of purity, safety and consistency have been addressed. A new
cell line (DCL, SCL or CCL) should be presumed to be tumorigenic unless data
132
Annex 3
demonstrate that it is not. If a manufacturer proposes to characterize the cell line as

non-tumorigenic, the following tests should be undertaken.
Cells from the MCB or WCB propagated to the proposed in vitro cell
age used for production, or beyond, should be examined for tumorigenicity, in a
test approved by the NRA/NCL. The test should involve a comparison between
the cell line and a suitable positive reference preparation (e.g. HeLa cells) and a
standardized procedure for evaluating results.
B.8.2 Type of test animals

A variety of animal systems have been used to assess the tumorigenic potential of
cell lines. Table A3.4 lists several examples of such tests, along with advantages
and disadvantages of each. Because assessing the tumorigenic phenotype of a
cell substrate requires the inoculation of xenogeneic or allogeneic cells, the test
animal should be rendered deficient in cytotoxic T-lymphocyte (CTL) activity.
This can be accomplished either by the use of rodents that are genetically
immunocompromised (e.g. nude mice, severe combined immunodeficiency
(SCID) mice) or by inactivating the T-cell function with antithymocyte globulin
(ATG), antithymocyte serum (ATS) or antilymphocyte serum (ALS). The
use of animals with additional defects in natural killer (NK)-cell function –
such as the non-obese diabetic (NOD)-SCID mouse, the NOD-SCID-gamma
mouse and the CD3 epsilon mouse – has not yet been explored for cell-substrate
evaluation, but they may offer some advantages. In addition to these systems,
several other in vivo systems such as the hamster cheek pouch (HCP) model
and ATG-treated non-human primates have been used in the past, but are rarely
used at present.
Table A3.4
In vivo tests to assess the tumorigenic potential of inoculated cells
Test and brief Advantage(s) Disadvantage(s) References

description
Adult athymic • Animals readily • Higher frequency of (65)
mouse (Nu/Nu available spontaneous tumours than
genotype): animals • No immuno- in other animal models
inoculated by the suppression that are not genetically
intramuscular or required immunosuppressed
subcutaneous • Low sensitivity for
routes with cells to assessing the metastatic
be tested potential of the inoculated
cells
133

description
Newborn athymic • No immuno- • Low sensitivity for assessing (66)
mouse: animals suppression the metastatic potential of
inoculated by the required the inoculated cells
subcutaneous route • More sensitive • Since litters include
with cells to be than adults heterozygous mice, twice
tested the number of animals
must be inoculated in order
to be sure that a sufficient
number of homozygous
mice have been included.
• Cannibalism of neonates by
the mother
SCID mouse: • No immuno- • Highly susceptible to (66, 67)

animals receive suppression viral, bacterial and fungal
subcutaneous, required infections
intradermal or • Potentially • Infections can affect the
intrakidney capsule increased sensi- results and reproducibility
inoculation of test tivity of studies
cells
• Animals readily • Spontaneous thymic
available lymphomas may occur
Newborn rat: animals • Animals readily • Standardized ATG not (65, 68)
immunosuppressed available available as a commercial
with ATG, followed • Sensitive model product
by intramuscular for detecting • Careful qualification and
or subcutaneous metastases characterization of the
inoculation of cells ATG is required to find

• Very low fre-
to be tested the balance between
quency of spon-
taneous tumour immunosuppressive
formation capacity and toxicity
Newborn hamster • Animals readily • Cannibalism of neonates (69)

or mouse: animals available by the mother
immunosuppressed • Standardized ATS not
with ATS, followed available and difficult to
by intramuscular balance toxicity versus
or subcutaneous immunosuppressive
inoculation of cells capacity
to be tested
134
Annex 3

description
Newborn hamster • Increased sensi- • Cannibalism of neonates (70)
or mouse: animals tivity compared by the mother
immunosuppressed to the HCP test • Standardized ALS not
with ALS, followed • Animals readily available and difficult to
by intramuscular available balance toxicity versus
or subcutaneous immunosuppressive
inoculation of cells capacity
to be tested
HCP: animals • Animals readily • Lower sensitivity than (71)
immunosuppressed available newer models
with cortisone,
followed by
inoculation of the
cells to be tested
into the cheek
pouch
Non-human • Species closer • Standardized ATG not (72)
primates: animals to human available
immunosuppressed • Animals not readily
with ATG, followed available
by inoculation of
• Expense and limited
cells to be tested
availability preclude using
into the muscle of
large numbers
the four limbs
• Animal welfare principles
mandate against use of
non-human primates
if the same results can
be obtained from lower
species
Although all the animal models listed in Table A3.4 have been used to
assess the tumorigenicity of cells, several sensitivity parameters from studies
using positive-control cells should be considered when attempting to compare
the various in vivo tumorigenicity models. These sensitivity parameters are:
(i) frequency of tumour formation;
(ii) time to appearance of tumours;
(iii) size of tumours;
135
(iv) lowest number of inoculated tumour cells that result in tumour

formation;
(v) metastatic tumour formation.
Factors (i), (ii), (iii) and (v) usually depend on the number of cells
inoculated (i.e. they are dose dependent). In addition, the rate of spontaneous
tumour formation should be considered. Although comparisons of two or more
assays have been reported in the literature (68, 73, 74) none take all of these
factors into account, nor do they use the same tumorigenic cell lines. Thus, it is
not possible to draw definitive conclusions about the relative sensitivity of the
various tumorigenicity assays. Nevertheless, the following points appear to be
generally accepted:
■ the ATS-treated newborn rat and the ATG-treated non-human-
primate systems are the most sensitive for assessing the metastatic
potential of inoculated cells;
■ ATS and ATG provide better immunosuppression than ALS;
■ the nude mouse has a better-defined level of immunosuppression
than models that depend on ALS, ATS or ATG, and interlaboratory
comparisons of data from nude mice are more likely to yield valid
conclusions.
Overall experience during the past 30 years, taking into consideration the
points mentioned above, has led to the conclusion that the athymic nude mouse
is an appropriate test system for determining the tumorigenic potential of cells
proposed for use in the production of biologicals. The major advantages of the
athymic nude mouse system are that it is easier to establish and standardize and
is generally available, while the newborn rat system is more sensitive for assessing
the metastatic potential of tumorigenic cells. In some cases, it may be preferable
to use newborn athymic nude mice, as these animals are more sensitive than
adults for the detection of weakly tumorigenic cells (66). A tumorigenicity testing
protocol using athymic nude mice is provided as Appendix 2. The animal system
selected should be approved by the NRA/NCL.
B.8.3 The point in cells’ life history at which they should be tested
Investigation of tumorigenicity should form part of the early evaluation of a new
cell substrate for use in production. Cells from the MCB or WCB, propagated
to the proposed in vitro cell age and used for production or beyond should be
examined for tumorigenicity. The extra population doublings (e.g. 3–10) ensure
that the results of the tumorigenicity test can be used in the assessment of overall
safety of the product, even assuming a worst-case situation, and this therefore
provides a safety buffer.
136
Annex 3
B.8.4 Use of control cells

The tumorigenicity test should include a comparison between the CCL and a
positive-control reference preparation such as HeLa cells from a reliable source.
This source is preferred in order to standardize the test between laboratories,
so that cumulative experience over time can be assessed and made available to
NRAs/NCLs and manufacturers, to assist them in the interpretation of data.
However, other sources for establishing positive-control cells may be acceptable.
The purpose of the positive control is to assure that an individual test is valid,
by demonstrating that the animal model has the capacity to develop tumours
from inoculated cells (i.e. a negative result is unlikely to be due to a problem
with the in vivo model). If the positive-control cells fail to develop tumours
at the expected frequency, then this could be indicative of problems (such as
infections) in the animals or in the testing facility, which can reduce the efficiency
of tumour development.
When the cell substrate has been adapted to growth in serum-free
medium, which may contain growth factors and other components that could
influence growth as well as detection of a tumorigenic phenotype, consideration
should be given to processing the positive-control cells in the same medium.
Whenever possible, both the test article and the positive-control cells should
be resuspended in the same medium, such as phosphate-buffered saline (PBS),
for inoculation.
In designing a tumorigenicity protocol, it is important to recognize that
tumours arise spontaneously in nude mice and that the incidence of such tumours
increases with the age of the mice. Therefore, databases (both published data and
the unpublished records/data of the animal production facility that supplied the
test animals) of rates of spontaneous neoplastic diseases in nude mice should be
taken into account during the assessment of the results of a tumorigenicity test. In
general, negative controls are not recommended because the rates of spontaneous
neoplastic disease in nude mice are low, and small numbers of negative-control
animals are unlikely to provide meaningful data. However, if negative-control cells
such as WI-38, MRC-5 or FRhl-2 are included, clear justification for including
them should be provided. For example, if serum-free medium is used to grow the
cell substrate, it is conceivable that growth factors may influence the appearance
of spontaneous tumours; consequently, negative-control cells suspended in the
same medium may be needed to interpret the test results.
B.8.5 Number of test animals

To determine whether the cells being characterized have the capacity to form
tumours in animals, the cells being tested, the reference positive-control cells and,
if any, the reference negative-control cells should be injected into separate groups
of 10 animals each. In a valid test, progressively growing tumours should be
produced in at least 9 out of 10 animals injected with the positive reference cells.
137
B.8.6 Number of inoculated cells

Each animal should be inoculated intramuscularly or subcutaneously (75) with
a minimum of 10 7 viable cells. If there is no evidence of a progressively growing
nodule at the end of the observation period, the cell line may be considered to
be non-tumorigenic. If the cell line is found to be tumorigenic, the NRA/NCL
may request additional studies to determine the level of tumorigenicity. This can
be done with dose–response studies, where doses of 10 7, 10 5, 10 3 and 10 1 viable
cells are inoculated, and the data can be expressed as tumour-producing dose at
the 50% end-point (TPD50 value) (76).
B.8.7 Observation period

Animals are examined weekly by observation and palpation for evidence of
nodule formation at the site of injection. The minimum observation period
depends on the test system selected. In the case of the nude mouse, a minimum
of 4 months is recommended. A shorter period is recommended for the ATS-
treated newborn rat because the immunosuppressive effect of the ATS declines
after the final injection at 2 weeks.
B.8.8 Assessment of the inoculation site over time

(progressive or regressive growth)
If nodules appear, they are measured in two perpendicular dimensions, the
measurements being recorded weekly to determine whether the nodule grows
progressively, remains stable or decreases in size over time. Animals bearing
nodules that are progressing should be killed before the end of the study if the
tumour reaches the limit set by the relevant authorities for the humane treatment
of animals. Animals bearing nodules that appear to be regressing should not be
killed until the end of the observation period. Cell lines that produce nodules that
fail to grow progressively are not considered to be tumorigenic. If a nodule persists
during the observation period and retains the histopathological characteristics of

a tumour, this should be investigated further and discussed with the NRA/NCL.
B.8.9 Final assessment of the inoculation site

At the end of the observation period, all animals, including the reference group(s),
are killed and examined for gross and microscopic evidence of the growth of
inoculated cells at the site of injection and other sites.
B.8.10 Evaluation of animals for metastases

Animals are examined for microscopic evidence of metastatic lesions in sites
such as the liver, heart, lungs, spleen and regional lymph nodes.
138
Annex 3
B.8.11 Assessment of metastases (if any)

Any metastatic lesions are examined further to establish their relationship to the
primary tumour at the injection site. If what appears to be a metastatic tumour
differs histopathologically from the primary tumour, it is necessary to consider
the possibility that this tumour either developed spontaneously or was induced by
one or more of the components of the cell substrate, such as an oncogenic virus.
This may require further testing of the tumour itself, or the tumorigenicity assay
may need to be repeated. In such cases, appropriate follow-up studies should be
discussed and agreed with the NRA/NCL (also see section B.9, “Oncogenicity”).
B.8.12 Interpretation of results

A CCL is considered to be tumorigenic if at least 2 out of 10 animals develop
tumours at the site of inoculation within the observation period. However, the
reported rate of spontaneous neoplastic diseases in the test animals should
be taken into account during the assessment of the results. In addition, the
histopathology of the tumours must be consistent with the inoculated cells, and
a genotypic marker should show that the tumour is not of nude mouse origin.
If only one of 10 animals develops a tumour, it is appropriate to investigate
further in order to determine, for example, if the tumour originated from the
cell substrate inoculum or from the host animal and whether there are any viral
or inoculated cell DNA sequences present. The NRA/NCL should be consulted
in this regard.
The dose–response of the CCL may be studied in a titration of the
inoculum as part of the characterization of the CCL. The need for such data
will depend on many factors specific to a given CCL and to the product being
developed. The NRA/NCL should be consulted in this regard.
Applicability
■ Cell banks: representative EOPC or ECB from the MCB or first WBC
B.9 Oncogenicity
B.9.1 Tests for oncogenicity
While tumorigenicity is the property of cells to form tumours when inoculated
into susceptible animals, oncogenicity is the property of an acellular agent to
induce cells of an animal to become tumour cells. As such, tumours that arise
in a tumorigenicity assay contain cells derived from the inoculated cells, while
tumours that arise in an oncogenicity assay are derived from the host. Oncogenic
activity from cell substrates could be due either to the cell substrate DNA (and
139
perhaps other cellular components) or to an oncogenic viral agent present in

the cells. Although there may be a perception that the cellular DNA from highly
tumorigenic cells would have more oncogenic activity than the DNA of weakly or
non-tumorigenic cells, it is not currently known if there is a relationship between
the tumorigenicity of a cell and the oncogenicity of its DNA. Nevertheless, the
NRA/NCL may require oncogenicity testing of the DNA and cell lysate from a
new cell line (i.e. other than those such as CHO, NS0, Sp2/0 and low-passage Vero,
for which there is considerable experience) that is tumorigenic in animal model
systems (see below), because of the perception that a vaccine manufactured in
such a cell line poses a neoplastic risk to vaccine recipients.
The major complication in assaying cellular DNA in animals arises from
the size of the mammalian genome. Because the mammalian haploid genome is
approximately 3 × 10 9 base pairs (bp), whereas the size of a typical oncogene could
be 3–30 × 10 3 bp, the concentration of an oncogene in cellular DNA expression
systems would be about 10 5–10 6-fold less concentrated than a plasmid containing
the same oncogene. As a consequence, if 1 µg of an oncogene expression plasmid
induces a tumour in an experimental animal model, the amount of cellular DNA
that would contain a similar amount of the same oncogene is 10 5–10 6 µg (i.e.
100 mg to 1 g). To date, three studies have indicated that between 1 and 10 µg
of expression plasmids for cellular oncogenes can be oncogenic in mice (34, 43,
44). Therefore, more sensitive in vivo assays need to be developed before the
testing of the oncogenic activity of cellular DNA becomes practicable. Recent
results suggest that the sensitivity of the assay can be increased by several orders
of magnitude, with the use of certain immune-compromised strains of mice that
are prone to develop tumours after inoculation with oncogenes. Thus, it may be
possible to assess the oncogenic activity of cellular DNA in the future. However,
at present there is no standardized in vivo oncogenicity test for cellular DNA.
An example of a protocol is nonetheless provided in Appendix 3.
Several in vitro systems, such as scoring the neoplastic transformation of
NIH 3T3 cells in a focus-forming assay following transfection of oncogenic DNA
(77–79), have been used to assess oncogenicity. However, it is not clear how such
assays reflect the oncogenic activity of DNA in vivo, since they predominantly
detect the oncogenic activity of activated ras-family members, and thus it is
unclear how the assays can assist in estimating risk associated with the DNA or
cell lysate from a cell substrate.
Based on experience with DCLs WI-38, MRC-5 and FRhL-2, testing of
new MCBs of these cell lines for oncogenicity is not recommended. Other DCLs
for which there is substantial experience may also not need to be tested. The NRA/
NCL should be consulted in this regard. As stated in section B.8.1, a new CCL
should be presumed to be tumorigenic unless data demonstrate that it is not. If
a manufacturer demonstrates that a new CCL is non-tumorigenic, oncogenicity
testing on cell DNA and cell lysates might not be required by the NRA/NCL.
140
Annex 3
When appropriate, and particularly for vaccines, cell DNA and cell lysates
should be examined for oncogenicity in a test approved by the NRA/NCL. An
oncogenicity testing protocol is provided as Appendix 3.
Applicability
■ Cell banks: MCB or first WCB taken to representative EOPC or ECB

■ Cell types: CCL, SCL (recommended when tumorigenic cells are used
in vaccine production)
B.10 Cytogenetics
B.10.1 Characterization
Chromosomal characterization and monitoring were introduced in the 1960s,
to support the safety and acceptability of human DCLs as substrates for vaccine
production. Human DCLs differ from CCLs by retaining the characteristics
of normal cells, including the normal human diploid karyotype. A significant
quantity of data has been accumulated since then, and this has led to the
conclusion that less extensive cytogenetic characterization is appropriate because
of the demonstrated karyotypic stability of human DCLs used in vaccine
production (80). Thus, the use of karyology as a lot-by-lot quality-control test is
unnecessary for well-characterized and unmodified human DCLs (e.g. WI-38,
MRC-5) and for FRhL-2.
Cytogenetic data may be useful for the characterization of CCLs,
especially when marker chromosome(s) are identified. Such data may be helpful
in assessing the genetic stability of the cell line as it is expanded from the MCB
to the WCB and finally to production cultures (see section B.3). The following
recommendations are appropriate for the characterization of DCL and CCL
cell banks.
Cytogenetic recharacterization of DCLs (e.g. WI-38, MRC-5 and FRhL-2)
should not be required unless the cells have been genetically modified or the
culture conditions have been changed significantly, since such data are already
available (19–21). However, for each WCB generated, manufacturers should
confirm once that the cells grown in the manner to be used in production are
diploid and have the expected lifespan.
To determine the general character of a new or previously uncharacterized
DCL, samples from the MCB should be examined at approximately four equally
spaced intervals during serial cultivation from the MCB through to the proposed
in vitro cell age used for production or beyond. The testing intervals should be
agreed with the NRA. Each sample should consist of a minimum of 100 cells in
metaphase and should be examined for exact counts of chromosomes as well as
for breaks and other structural abnormalities.
141
Giemsa-banded karyotypes of an additional five metaphase cells in each

of the four samples may provide additional useful information. The
ISCN (81) 400 band is the minimum acceptable level of Giemsa-banding
analysis for human cells.
Stained slide preparations of the chromosomal characterization of the

cells (i.e. DCL, CCL), or photographs of these, should be maintained permanently
as part of the cell-line record. Further recommendations have been proposed for
SCLs (57).
Applicability
■ Cell banks: MCB, ECB or representative EOPC

■ Cell types: DCL, SCL, CCL (as a test for genetic stability, when
appropriate)
B.11 Microbial agents

B.11.1 General considerations
While many biological production systems require human or animal cell
substrates, such cells are subject to contamination with, and have the capacity
to propagate, extraneous, inadvertent or so-called adventitious organisms such
as mycoplasma and viral agents. In addition, animal cells contain endogenous
agents such as retroviruses that may also be of concern. Testing for both
endogenous (e.g. retroviruses) and adventitious agents (e.g. mycoplasmas) is
described in the subsequent sections. In general, cell substrates contaminated
with microbial agents are not suitable for the production of biologicals. However,
there are exceptions to this general rule. For example, the CHO and other rodent
cell lines that are used for the production of highly purified recombinant proteins
express endogenous retroviral particles. The balance of risk versus benefit must be
considered when determining the suitability of a cell substrate for the production
of a specific product. Further, risk-mitigation strategies during production,
including purification (removal) and inactivation by physical, enzymatic and/or
chemical means, should be implemented whenever appropriate and feasible. Even
though a cell substrate might be unacceptable for some products, such as a live
viral vaccine subjected to neither significant purification nor inactivation, that
same cell substrate may be an acceptable choice for a different type of product,
such as a highly purified recombinant protein or monoclonal antibody for which
risk mitigation has been achieved by significant and validated viral clearance in
the production process.
A strategy for testing cell banks for microbial agents should be developed.
One strategy is to perform exhaustive testing at the MCB level and to carry out
142
Annex 3
more limited testing on the WCB derived from the MCB. This more limited
testing would be selected on the basis of those agents that could potentially be
introduced during the production of the WCB from the MCB. Testing would not
need to be replicated for agents that could only have been present prior to the
production of the MCB (e.g. an endogenous retrovirus, or BVDV from serum
used for developing the cell seed or in the legacy of establishing the cell line).
However, if the number of vials of an MCB is limited, an alternative
strategy would be to conduct the more exhaustive testing on the first WCB
made from that MCB, and to limit testing on the MCB itself. An advantage to
the strategy of performing more exhaustive testing on the first WCB is that it
provides a greater opportunity for amplification of any agents that may have
been introduced earlier and through to production of the WCBs. There are
advantages and disadvantages to more extensive testing of the MCB or the WCB,
and consideration should be given to what is more appropriate for the particular
product(s) to be manufactured using a given cell bank. Consultation with the
NRA/NCL should be considered prior to implementation, to determine whether
a proposed testing strategy is acceptable.
EOPC/ECB should be characterized once for each commercial production
process. Testing of the ECB serves as further characterization of the MCB or
WCB that was exhaustively tested. It also permits additional time/passages for
amplification of low-level contaminants or reactivation of viral contaminants
that may have been missed in the testing of the upstream bank.
B.11.2 Viruses
Manifestations of viral infections in cell cultures vary widely among the broad
array of virus families that are potential contaminants; thus, the methods used to
detect them vary. Lytic infections are frequently detected by the CPE they cause.
However, in some cases such as non-cytopathic BVDV, no CPE is observed.
Viruses also may be present latently (e.g. herpesviruses) or endogenously (in
the germline, e.g. retroviral proviruses). Specific techniques such as molecular
and immunological methods and electron microscopy may be required to reveal
the presence of such inapparent infections. For new cell substrates, induction of
a detectable infection by exposing the cells to special conditions (e.g. chemical
induction, heat shock) may be required, and special detection techniques such as
transcriptome sequencing or degenerate primer PCR may be useful.
The strategy developed to test cell substrates for viruses should take into
consideration the families of viruses and specific viruses that may be present in
the cell substrate. Consideration should be given to the species and tissue source
from which the cell substrate originated, and to the original donor’s medical
history in the case of human-derived cell substrates or to the pathogen status of
donor animals in the case of animal-derived cell substrates. Consideration should
143
also be given to viruses that could contaminate the cell substrate from the donors
or from animal- or human-derived raw materials used in the establishment and
passage history (legacy) of the cell substrate prior to and during cell banking or
production (e.g. serum, trypsin, animal- or human-derived medium components,
antibodies used for selection, or animal species through which the cell substrate
may have been propagated), as well as laboratory contamination from operators
or other cell cultures.
Tests should be undertaken to detect, and where possible identify, any
endogenous or exogenous agents that may be present in the cells. Attention
should be given to tests for agents known to cause an inapparent infection in the
species from which the cells were derived, thereby making it more difficult to
detect (e.g. simian virus 40 (SV40) in rhesus monkeys).
Primary cells are obtained directly from the tissues of healthy animals and
are more likely to contain adventitious agents than banked, well-characterized
cells. In addition, recent vaccination of source animals should be considered, as
the animals may be exposed to live vaccines. The risk with primary cells can be
mitigated by rigorous qualification of source animals and of the primary cells
themselves. When feasible, animals from which primary cultures are established
should be from genetically closed flocks, herds or colonies that are monitored
for freedom from pathogens of specific concern. Such animals are known as
specific-pathogen-free, or SPF. The term “closed” refers to the maintenance of
a group (flock, herd or colony) free from introduction of new animals (new
genetic material that could introduce new retroviral proviruses, for instance).
Many live viral vaccines are commonly produced in primary cells and undergo
little purification during production. In such cases, and when feasible, the use of
SPF animals is highly recommended. Documentation of the status of the source
animals should be provided to the NRA/NCL. Animals that are not from closed
flocks, herds or colonies should be quarantined and thoroughly evaluated for a
period that is sufficient to detect signs of disease or infection (e.g. monkeys are
generally quarantined for six or more weeks (82)). Such animals also should be
screened serologically for appropriate adventitious agents, in order to determine
their suitability as a source for the primary cell substrate. Animal husbandry
practices should be documented. Even so, viral contamination of the cells may
not be excluded from all cultures. For example, contamination of primary
monkey kidney cells with foamy virus or simian cytomegalovirus is common in
the absence of specific concerted efforts to prevent such contaminations.
For primary cell cultures, the principles and procedures outlined in Part C
of Recommendations for the production and control of poliomyelitis vaccine
(oral) (82), together with those in section A.4 of Requirements for measles,
mumps and rubella vaccines and combined vaccine (live) (83) may be followed.
The production of viral vaccines, such as those against smallpox or rabies,
originally required the use of living animals, and the great range of possible viral
144
Annex 3
contaminants became apparent only as cell culture methods were developed.

For example, human enteroviruses were not recognized until the development
of monkey kidney cell cultures in which they could produce CPEs, because
the disease produced in humans is either relatively mild or in some cases non-
existent. It was also clear that viruses could be detected in some systems but not
others. For instance, the polyomavirus SV40 does not produce a CPE in cultures
from rhesus monkey kidney cells (in which much of the early polio vaccines were
produced) derived from SV40-infected monkeys but will do so in cultures from
cynomolgus or African green monkey kidney cells. The suspicion was, therefore,
that there were many viruses in the culture systems of the time and that they were
detected only if the assays were appropriate. This remains an accurate view and
has led to a range of different approaches for trying to detect all contaminants.
Coxsackie viruses are named after the town in New York where they were
first identified after being detected because of their effects in mice. Coxsackie B
viruses produce clinical signs and death in adult mice, while Coxsackie A viruses
will affect only suckling mice. For many years, tissue culture methods were a less
reliable method of detection of Coxsackie A viruses than suckling mice, and the
continued use of these animals in cell bank characterization reflects this.
In the 1940s, embryonated chicken eggs were a popular substrate for the
growth and assay of viruses such as influenza, measles, mumps, yellow fever and
vaccinia. They therefore appear to have a wide range of susceptibility. Simian
viruses such as SV5 or viruses such as Sendai virus also grow well in them.
As many are paramyxoviruses with haemagglutinating activity, the egg-based
assays include tests for haemagglutinating activity, as well as for the death of
the embryos.
A range of tissue culture cells is also used, typically including one human,
one of the same species as the production cell, and one other (often of monkey
origin). The hope is that the range will catch viruses not detected by other means,
although in practice it is wise to assume that there is no such thing as a generic
detection method; for example, cells from an inappropriate monkey species
will not necessarily detect SV40, whereas cells from other species will (e.g. Vero
cells). In certain circumstances where a virus is of particular concern, specific
tests have been applied. For example, herpes B virus is a common infection in
monkeys in the absence of precautions such as quarantine and clinical evaluation
of the donor animals, and this has very serious effects on infected humans. While
herpes B virus was routinely detected by the use of primary rabbit kidney cell
cultures, established rabbit cell lines are now acceptable for this purpose. Another
example is Marburg virus, which in the 1970s caused a number of deaths in
workers who handled monkeys that were to be used in a vaccine production
facility. The incident might have been avoided if the animals had been adequately
quarantined. A specific test in guinea-pigs was introduced and maintained for a
number of years to ensure the absence of the agent.
145
There is a disparate range of tests that have been or still are used with the
aim of detecting any significant contaminant that may be present in cell cultures.
Some, such as the rabbit kidney cell test, are very specific in intent, while others
may be expected to be more generic. In general, however, it is wise to assume that
an assay will never be all-encompassing, whether based on historical virological
approaches or more current methodologies. Thus, the consequences of deleting
tests on the grounds of redundancy must be very carefully evaluated before any
action is taken. On the other hand, it is difficult to justify the maintenance of a
test if it detects only viruses that are also detected by other methods of equivalent
sensitivity and comparable ease of use and cost. Each of these considerations
should be borne in mind when developing an appropriate testing strategy for the
given cell bank. Policies to minimize the use of animals in safety testing should
also be considered, but must be balanced with the utility and necessity (sensitivity
and ability to detect particular adventitious agents not readily detected by other
means) of the test in which they are used.
B.11.2.1 Tests in animals and eggs

The cells of the MCB and WCB are unsuitable for production if any of the animal
or egg tests show evidence of the presence of any viral agent attributable to the
cell banks.
In general, MCBs are thoroughly characterized by the methods listed
below. WCBs may be characterized by an abbreviated strategy, when appropriate.
However, an alternative strategy to this general rule may be used, as discussed in
section B.11.2 above. These tests may be performed directly on cells, supernatant
fluids or cell lysates from the bank itself, or on cells or supernatant fluids or cell
lysates from cells from the bank that have been passaged to the proposed in vitro
cell age for production or beyond.
In some countries, policies exist to minimize the use of animals in
safety testing.
B.11.2.1.1 Adult mice

The original purpose of this test was to detect lymphocytic choriomeningitis
virus (LCMV). The test in adult mice for pathogenic viruses includes inoculation
by the intraperitoneal route (0.5 ml) with cells and culture fluids from the MCB
or WCB, where at least 10 7 viable cells or the equivalent cell lysate are divided
equally among at least 10 adult mice weighing 15–20 g.
In some countries, the adult mice are also inoculated by the intracerebral
route (0.03 ml).
In some countries, at least 20 mice are required for each test.

146
Annex 3
The animals are observed for at least 4 weeks. Any animals that are sick or that
show any abnormality are investigated to establish the cause. Animals that do
not survive the observation period should be examined for gross pathology and
histopathology, in order to determine a cause of death and, if a viral infection is
indicated, efforts should be undertaken to identify the virus. Viral identification
may involve culture and/or molecular methods. Further, each mouse that
dies after the first 24 hours of the test, or is killed because of illness, should be
necropsied and examined for evidence of viral infection by subinoculation of
appropriate tissue into at least five additional mice, which should be observed
for 21 days. The test is not valid if more than 20% of the animals in either the
test group or the negative control group (if used), or in both, become sick for
nonspecific reasons and do not survive the observation period.
In some countries, the adult mice are observed for 21 days.
If the cell substrate is of rodent origin, at least 10 6 viable cells or the equivalent
cell lysate are injected intracerebrally into each of 10 susceptible adult mice to test
for the presence of LCMV.
In some countries, after the observation period, the animals are challenged
with live LCMV to reveal the development of immunity against non-
pathogenic LCMV contaminants resulting in otherwise inapparent
infection.
Applicability
■ Cell banks: MCB, WCB or ECB or representative EOPC

■ Cell types: PCC, DCL, SCL, CCL
B.11.2.1.2 Suckling mice

The original purpose of this test was to detect Coxsackie viruses. The test for
pathogenic viruses in suckling mice includes inoculation by the intraperitoneal
route (0.1 ml) with cells and culture fluids from the MCB or WCB. At least 10 7
viable cells or the equivalent cell lysate are divided equally between two litters of
suckling mice, comprising a total of at least 10 animals less than 24 hours old.
In some countries, the suckling mice are also inoculated by the
intracerebral route (0.01 ml).
In some countries, 20 suckling mice are inoculated.
The animals are observed for at least 4 weeks. Any animals that are sick or
show any abnormality are investigated to establish the cause. Animals that do
not survive the observation period should be examined for gross pathology and
147
histopathology, in order to determine the cause of death. If a viral infection is

indicated, efforts should be undertaken to identify the virus where practicable.
Viral identification may involve culture and/or molecular methods. Further,
such examination of viral infection should include subinoculation of appropriate
tissue suspensions into an additional group of at least five suckling mice, by
intracerebral and intraperitoneal routes, and observation daily for 14 days. In the
case of suckling mice, it is often observed that those that perish are cannibalized
by their mother and this renders determination of cause of death impossible
(when they are fully cannibalized and no remains can be recovered). The test is
not valid if more than 20% of the animals in either the test group or the negative
control group (if used), or in both, do not survive the observation period.
In some countries, the suckling mice may be observed for a period
of 14 days, followed by a subpassage involving a blind passage (via
intraperitoneal and intracerebral inoculation into at least five additional
mice) of a single pool of the emulsified tissue (minus skin and viscera) of
all mice surviving the original 14-day test.
Applicability

B.11.2.1.3 Guinea-pigs
The original purpose of this test was to detect LCMV and Mycobacterium
tuberculosis. When it is necessary to detect Mycobacterium species, a test in
guinea-pigs is performed and includes inoculation by the intraperitoneal route
(5 ml) with cells and culture fluids from the MCB or WCB, where at least 10 7
viable cells or the equivalent cell lysate are divided equally among the animals.
In some countries, five guinea-pigs weighing 350–450 g are also

inoculated by the intracerebral route (0.1 ml) and observed for 42 days
to reveal Mycobacterium tuberculosis and other species.
The animals are observed for at least 6 weeks. Animals that are sick or show
any abnormality are investigated to establish the cause. Animals that do not
survive the observation period should be examined for gross pathology and
may involve culture and/or molecular methods. The test is not valid if more than
20% of the animals in either the test group or the negative control group (if used),
or in both, do not survive the observation period.
148
Annex 3
The test in guinea-pigs for the presence of Mycobacterium may be replaced

by an alternative in vitro method such as culture, or shortened culture with a
PCR end-point (also see section B.11.3).
Applicability

■ Cell types: PCC, DCL, SCL, CCL (the latter three are dependent on
legacy and current use of media components of animal origin that
could result in contamination with mycobacterial species)
B.11.2.1.4 Rabbits
The original purpose of this test was to detect herpes B virus. When it is necessary
to detect simian herpes B virus, the test in rabbits for pathogenic viruses is
performed and includes inoculation by the intradermal (1 ml) and subcutaneous
(>2 ml) routes with cells and culture fluids from the MCB or WCB, where at least
10 7 viable cells or the equivalent cell lysate are divided equally among the animals.
In some countries, five rabbits weighing 1.5–2.5 kg are inoculated by the
subcutaneous route, with either 2 ml or between 9 and 19 ml. Consultation
with the NRA/NCL regarding acceptable methods should be considered.
The animals are observed for at least 4 weeks. Animals that are sick or show
any abnormality are investigated to establish the cause. Animals that do not
survive the observation period should be examined for gross pathology and
may involve culture and/or molecular methods. The test is not valid if more than
20% of the animals in either the test group or the negative control group (if used)
do not survive the observation period.
The test in rabbits for the presence of herpes B virus is intended for
primary simian cultures, and may be replaced by a test in rabbit kidney cell
cultures.
Applicability

B.11.2.1.5 Embryonated chicken eggs

At least 10 6 viable cells or the equivalent cell lysate, along with culture fluids from
the MCB or WCB of avian origin, propagated to the proposed in vitro cell age
for production or beyond, are injected into the allantoic cavity of each of at least
149
10 embryonated hens’ eggs, and into the yolk sac of each of at least another 10
embryonated hens’ eggs. The eggs are examined after not fewer than 5 days of
incubation. The allantoic fluids of the eggs are tested with red cells from guinea-
pig and chicken (or other avian species) for the presence of haemagglutinins.
The test is not valid if more than 20% of the embryonated hens’ eggs in either the
test group or the negative control group (if used), or in both, are discarded for
nonspecific reasons.
In some countries, the NRA/NCL also requires that other types of red
cells, including cells from humans (blood group IV O) or monkeys,
should be used in addition to cells from guinea-pig and chicken (or other
avian species). In all tests, readings should be taken after incubation for
30 minutes at 0–4 °C, and again after a further incubation for 30 minutes
at 20–25 °C. For the test with monkey red cells, readings also should be
taken after a final incubation for 30 minutes at 34–37 °C.
In some countries, inoculation by the amniotic route is used.
In some countries, following incubation, allantoic fluids or a 10%

suspension of yolk sacs, as appropriate, should be harvested, pooled and
blind passaged into an additional group of eggs.
The eggs used for the yolk-sac test should usually be 5–7 days old. The eggs used
for the allantoic cavity test should be 9–11 days old.
Alternative ages for the embryonated chicken eggs and alternative
incubation periods are acceptable if they have been determined to be
equivalent or better for detecting the presence, in the test samples, of the
adventitious agents that the test is capable of detecting when performed
as above.
Embryos that do not survive the observation period should be examined for
gross pathology, in order to determine a cause of death and, if a viral infection is

may involve culture and/or molecular methods.
Applicability
■ Cell banks: MCB of avian origin, WCB of avian origin or ECB or

representative EOPC
■ Cell types: avian PCC, DCL, SCL, CCL (also recommended for novel
cell substrates)
B.11.2.1.6 Antibody production tests

Rodent cell lines are tested for species-specific viruses using mouse, rat and
hamster antibody production tests, as appropriate. In vivo testing for lymphocytic
150
Annex 3
choriomeningitis virus, including a challenge for non-lethal strains, is performed

for such cell lines, as described in section B.11.2.1.1. Avian cell lines may also be
tested using a chicken antibody production test – e.g. to detect chicken anaemia
virus. Further, if the cell substrate (even if not of rodent origin) has been exposed
to materials of rodent origin (e.g. selection using a monoclonal antibody),
testing should be considered for the species-relevant viruses, using an antibody
production test (84, 85).
In some countries, consideration is being given to use of nucleic acid
testing in place of the in vivo antibody production testing. In these cases,
data should be provided to the NRA/NCL to justify this practice.
Applicability

■ Cell types: DCL, SCL, CCL (recommended primarily for cells of
rodent origin)
B.11.2.2 Tests in cell culture

Tests in cell culture are capable of detecting a broad array of viral families.
Readouts include monitoring the cultures periodically for CPE and tests for
haemadsorbing and haemagglutinating viruses, which are conducted at the end
of the culture period. In addition to the indicator cells described below, it may
be appropriate to expand the different types of indicator cells used (beyond two
or three) to enable the detection of viruses with differing host requirements.
Decisions about which cell lines to use as indicator cells should be guided by the
species and legacy of the production cell substrate, taking into consideration the
types of viruses to which the cell substrate could potentially have been exposed
and thus the viruses one would like to detect by this assay method. The cell
substrate is unsuitable for production if any of the indicator cell cultures shows
evidence of the presence of any viral agent attributable to the tested cell substrate.
Applicability

B.11.2.2.1 Indicator cells

Live cells or cell lysate, each with spent culture fluids of the MCB or WCB, are
inoculated on to monolayer cultures or cultivated with monolayer cultures of the
cell types listed below, as appropriate.
151
A lysate of the cells may be prepared by a method that avoids virus

disruption while allowing maximal virus release (e.g. typically three
freeze/thaw cycles followed by low-speed centrifugation). If cells, lysate
or spent culture fluids are to be stored prior to testing, they should be
stored at ≤–70 °C.
Cultures (primary cells or CCL) of the same species and tissue type as that used
for production may be used.
Cultures of a human DCL may be used. The original purpose of this
test, using primary human cells, was the detection of measles virus. Where the
cell substrate is of human origin, a simian kidney cell line should be used as the
second indicator cell line. The original purpose of the use of this cell type was
the detection of simian viruses.
In some countries, cultures of another (third) cell line from a different
species are required.
In many circumstances, more than two cell lines may be necessary to

cover the range of potential viral contaminants and, typically, a third
cell line of simian origin would be used if the cell substrate is not of
simian origin.
For new cell substrates, additional cell lines to detect viruses known to be
potentially harmful to humans could be considered (e.g. for insect cell lines; if
the cells selected for the above-mentioned tests are not known to be permissive to
insect viruses, an additional detector cell line should be included in the testing).
The cell bank sample to be tested is diluted as little as possible. At
least 10 7 cells, or equivalent cell lysate, and spent culture fluids are inoculated
on to each of the indicator cell types. The resulting co-cultivated or inoculated
cell cultures are observed for evidence of viruses by CPE for at least 2 weeks.
If the cell line is known to be capable of supporting the growth of human or
simian cytomegalovirus, HDC cultures are observed for at least 4 weeks.
Extended (4 week) cell culture for the purposes of detecting human or simian
cytomegalovirus can be replaced by the use of NAT to detect cytomegalovirus
nucleic acid.
In some countries, a passage on to fresh cultures for an additional 2 weeks
is recommended for all indicator cultures. In some cases, it may be
difficult to keep the cell cultures healthy for 2 weeks without subculturing.
In these cases, it may be necessary to feed the cultures with fresh medium
or to subculture after 2 weeks on to fresh cultures, in order to be able to
detect viral agents.
At the end of the observation period, samples of each of the co-cultivated

or inoculated cell culture systems are tested for haemadsorbing and/or
haemagglutinating viruses, as described in section B.11.2.1.5.
152
Annex 3
B.11.2.2.2 Additional considerations regarding the tests in cell culture for insect viruses
Many insect cell lines carry persistent viral infections that do not routinely
produce a noticeable CPE (e.g. some clones of the Hi-5 cell line are persistently
infected with an insect nodavirus). However, the viruses may be induced to
replicate by stressing the cells with a variety of techniques such as increased/
reduced culture temperature (above or below that routinely used for production),
heat shock for a short period, superinfection with other insect viruses, or chemical
inducers. Therefore, the probability of detecting such low-level persistent
infections may be increased by stressing the cells prior to analysis.
Intact cells and cell lysates from a passage level at or beyond that equivalent
to the EOPC are co-cultivated with indicator cells from at least three different
species of insect in addition to the indicator cells noted in section B.11.2.2.2.
Cell lines should be selected on the following basis: one of the lines has been
demonstrated to be permissive for the growth of human arboviruses, a second
line has been shown to be permissive for the growth of a range of insect viruses,
and the third has been derived from a species that is closely related to the host
from which the MCB is derived (or another line from the same species). Duplicate
cultures of indicator cells are typically incubated at two temperatures – such as
37 ± 1 °C and a lower temperature such as 28 ± 1 °C – observed for a period of
14 days, and examined for possible morphological changes. The cell culture fluids
from the end of the test period are tested for haemagglutinating viruses, or the
intact cells from the end of the test period are tested for haemadsorbing viruses.
The cells comply with the test if no evidence of any viral agent is found.
Several mosquito cell lines are available that are permissive for the growth
of some human arboviruses and could be considered for these tests. Alternatively,
BHK-21 cells could be considered for this purpose. The most permissive insect
cell lines characterized to date have been derived from embryonic Drosophila
tissues. While the mosquito and Drosophila cell lines may be suitable for some
aspects of the testing, it should be remembered that many insect cell lines are
persistently infected with insect viruses that usually produce no obvious CPE.
In addition, many insect cells may be infected with mammalian viruses, such
as BVDV, that are known to replicate in insect cells. Demonstrating that the
indicator cell lines are themselves free from adventitious agents is an important
prerequisite to their use in the testing outlined above. Consideration should also
be given to risk-mitigation strategies, as discussed above, for highly purified
products for which viral clearance can be achieved and validated.
B.11.2.3 Transmission electron microscopy

At least 200 cells from the MCB or WCB and from the ECB are examined
by transmission electron microscopy (TEM) for evidence of contamination
153
with microbial agents. Methods include negative staining and thin section. A
discussion of these methods is provided by Bierley et al. (86). In some cases it
may be appropriate to examine more cells, as discussed below for insect cell lines.
The NRA/NCL should be consulted in this regard. Any unusual or equivocal
observations that may be of microbiological significance should be noted and
discussed with the NRA/NCL.
TEM can detect viral particles in a cell substrate, including certain
endogenous retroviruses. While TEM is fairly insensitive (generally detecting
gross contamination, but not necessarily low-level contamination), it is a generic
assay that can detect microbial agents of many types.
Applicability
■ Cell banks: MCB, WCB, or ECB or representative EOPC

B.11.2.3.1 Additional considerations on TEM for insect cells

The general screening test outlined above applies to MCBs and WCBs derived
from insect cells. In addition, cell lines should be subjected to stress conditions,
such as described in section B.11.2.2.3, prior to examination by TEM. Increasing
the number of cells examined may also improve the probability of detecting an
agent (e.g. errantiviruses and hemiviruses). The maintenance temperatures and
treatments used should be agreed with the NRA/NCL, as should the number of
sectioned cells to be examined.
B.11.2.4 Tests for retroviruses

All vertebrate and insect cells that have been analysed possess endogenous,
genetically acquired retroviral sequences integrated into chromosomal DNA
in the form of proviruses. These sequences may be expressed, or may be
induced, as mRNA. In some cases, the mRNA is translated into viral protein,
and virus particles (virions) are produced. In many cases, these virions are
defective for replication (e.g. avian endogenous retrovirus EAV (endogenous
avian retrovirus), CHO cell line gamma-retrovirus (87)), whereas in others
(e.g. X-MuLV) the retroviruses may be capable of infecting cells of other species,
including human cells.
Consideration should also be given to the possibility that cell banks may
be infected with non-genetically acquired retroviruses (exogenous retroviruses),
either because the donor animal was infected or through laboratory contamination.
It should be noted that infection by retroviruses is not necessarily
associated with any CPE on the cells. Therefore screening assays, such as the PERT
assay for reverse transcriptase, or TEM may be required to reveal their presence.
154
Annex 3
The cells of the MCB or WCB are unsuitable for production if the
tests for infectious retroviruses, if required, show evidence of the presence of
any viral agent attributable to the substrate that cannot be demonstrated to be
cleared during processing. Generally, the downstream manufacturing process
for products (e.g. monoclonal antibodies) made in cell substrates that produce
retroviral particles (e.g. CHO cells) or infectious endogenous retrovirus (i.e. NS0,
Sp2/0 cells) is validated to provide adequate viral clearance (14). The margin of
viral clearance required should be agreed with the NRA/NCL.
Chick embryo fibroblasts (CEF) contain defective retroviral elements that
frequently produce defective particles with reverse transcriptase activity. This
has been the subject of many studies and WHO consultations because they are
used for production of live viral vaccine. If evidence is presented that the donor
flock is free of infectious retroviruses and there is no evidence that the cultures
are contaminated with infectious retroviruses, the cultures can be considered
acceptable with respect to retrovirus tests.
Rodent cell lines express endogenous retroviruses, and thus infectivity
tests should be performed to determine whether these endogenous retroviral
particles are infectious.
Cell lines such as CHO, BHK-21, NS0 and Sp2/0 have frequently been
used as substrates for drug production, with no reported safety problems
related to virus contamination of the products, and may be classified as “well
characterized” because the endogenous retrovirus particles have been studied
extensively. Furthermore, the total number of retrovirus-like particles present
in the harvest is evaluated quantitatively (TEM or quantitative PCR) on a
representative number of lots, and retrovirus clearance is demonstrated with
significant safety factors. In these situations, testing for infectious retrovirus may
be reduced (e.g. test one lot, then discontinue testing, but repeat when there is a
significant change in the cell culture process, such as a change in scale). Sponsors
are encouraged to consult with the NRA.
Applicability
MCBs and cells that have been propagated to the proposed in vitro cell age for
production or beyond. Alternatively, this testing could be performed on WCBs.
■ Cell banks: MCB or WCB, and ECB or representative EOPC
B.11.2.4.1 Reverse transcriptase assay

Test samples from the MCBs or WCBs propagated to the proposed in vitro cell
age for production or beyond are examined for the presence of retroviruses.
Culture supernatants are tested by a highly sensitive, quantitative PCR-
based reverse transcriptase (RT) assay or PERT assay (88–91).
155
RT activity is not specific to retroviruses and may derive from other

sources such as retrovirus-like elements that do not encode a complete or
infectious genome (92–96) or cellular DNA-dependent DNA polymerases
(97, 98). Attempts to reduce the PERT activity associated with cellular DNA-
dependent DNA polymerases have been reported (98–100), although no
treatment can eliminate all activity. Thus, the results of such highly sensitive
assays need to be interpreted with caution. Use of appropriate controls in the
assay can assist in this regard. Since RT activity can be associated with the
presence of defective retrovirus-like particles, and since polymerases other than
RT can result in apparent RT activity, a positive result in an RT assay is not
conclusive evidence of the presence of infective retrovirus. Positive results may
require further investigation, such as carrying out infectivity assays (see section
B.11.2.4.4). It may also be useful to utilize the conventional RT assay in this
investigation to determine whether the RT activity is Mg 2+ or Mn2+ dependent.
Such testing should be agreed in advance with the NRA/NCL.
CEFs and other cells of avian origin are known to express retroviral
elements. The appropriateness of this test with such cells should be discussed
with the NRA/NCL. For example, it may be appropriate to direct testing strategies
to the detection of infectious avian retroviruses, such as avian leukosis viruses
and reticuloendotheliosis virus, including serological screening of flocks that are
the source of the CEFs. Additionally, it is known that insect cells have retroviral
elements that are detected by a PERT assay, and so they too may test positive by
this assay.
B.11.2.4.2 PCR or other specific in vitro tests for retroviruses

If the PERT test gives unclear results, or when it is unavailable, it may be appropriate
to screen the cell substrate for species-specific retroviruses, by molecular methods
such as PCR, immunofluorescence, enzyme-linked immunosorbent assay (ELISA)
or other virus-specific detection methods. Molecular methods, such as PCR,
may also be used for quantification of retrovirus-like particles in the production
harvests, provided that the method is validated accordingly. Consultation with

the NRA/NCL regarding the acceptability of this approach is recommended.
B.11.2.4.3 Infectivity test for retroviruses

When the test sample is found to have RT activity, it may be necessary to carry
out infectivity assays to assess whether the activity is associated with replicating
virus.
Because rodent cells generally express endogenous retroviruses, the
infectivity and in vitro host range of such retroviruses should be assessed. Test
samples from the MCB or WCB, propagated to the proposed in vitro cell age for
production or beyond, should be examined with infectivity assays for the presence
of retroviruses. Cells to be used for these assays should be able to support the
156
Annex 3
replication of a broad range of viruses; this may require the use of cells of various
species and cell types. The testing strategy should be agreed with the NRA/NCL.
It is often possible to increase the sensitivity of assays by first inoculating
the test material on to cell cultures that can support retroviral growth, in order to
amplify any retrovirus contaminant that may be present at low concentrations.
For non-murine retroviruses, test cell lines should be selected for their capacity to
support the growth of a broad range of retroviruses, including viruses of human
and non-human primate origin (101, 102).
For murine retroviruses, it is important to assess whether the cells
release infectious retroviruses and, if so, to determine the host range of those
viruses. The testing for murine retroviruses can be complex, and the NRA/NCL
should be consulted for guidance. Murine and other rodent cell lines (CHO,
NS0, Sp2/0), or hybrid cell lines containing a rodent component, should be
assumed to be inherently capable of producing infectious retroviruses or non-
infectious retrovirus-like particles. In such cases, the clearance (removal and/or
inactivation) of such retroviruses during the manufacturing process should be
quantified and should provide a level of clearance acceptable to the NRA/NCL.
Any testing proposed by the manufacturer should be agreed with the
NRA/NCL.
B.11.2.5 Tests for particular viruses not readily detected by the tests
described in sections B.11.2.1–B.11.2.4 and their subsections
Some viruses, such as hepatitis B or C viruses or human papillomaviruses, cannot
be detected readily by any of the methods described above because these viruses
are not known to grow readily in cell culture, or are restricted to human host
range. Some animal viruses (e.g. bovine polyomavirus and porcine circoviruses)
are not readily detected by the routine tests previously described. In such
circumstances, it may be necessary to include specific assays for such viruses.
While broad general tests are preferable for detecting unknown contaminants,
some selected viruses may be screened by using specific assays such as molecular
techniques (e.g. nucleic acid amplification). Antibody-based techniques such as
immunofluorescence assays may also be employed.
Generally, once the MCB, WCB or ECB has been demonstrated to be free
of selected viruses, it may not be necessary to test the cells at later stages (e.g. at
the production level) if such viruses cannot be introduced readily during culture.
Human cell lines should be screened using appropriate in vitro techniques
for specific viruses that are the cause of significant morbidity, for those viruses
that might establish latent or persistent infections, and for viruses that may be
difficult or impossible to detect by the techniques described in sections B.11.2.1–
B.11.2.4 and their subsections. Selection of the viruses to be screened should take
into account the tissue source and medical history of the donor, if available, from
whom the cell line was derived.
157
Under circumstances in which the cell origin or medical history of the

donor, if available, would suggest their presence, it may be appropriate to perform
specific testing for the presence of human herpesviruses, human retroviruses,
human papillomaviruses, human hepatitis viruses, human polyomaviruses, or
difficult-to-culture types of human adenoviruses.
Consideration should be given to screening insect cell lines for specific
viruses that have been reported to contaminate particular cell lines (e.g.
nodaviruses) or that may be present persistently in insect cell lines and that are
known to be infectious for humans.
Applicability
The NRA/NCL should be consulted with regard to the specific pathogens or
selected viruses that should be included in the testing strategy, as these will be
directed on a case-by-case basis depending on the species and origin of the cell
and the medical history of the donor, if available.
■ Cell banks: MCB, WCB, or ECB or representative EOPC
■ Cell types: PCC (as needed), DCL, SCL, CCL
B.11.2.5.1 Nucleic acid detection methods

Tests for selected viruses are usually performed using nucleic acid amplification
and detection methods. PCR can be performed directly on DNA extracted from
the cells, or on cell lysates or supernatant fluids by DNA amplification, or on
RNA by reverse transcription followed by DNA amplification (RT-PCR). In this
manner, both DNA and RNA viruses can be detected, as can the proviral DNA of
retroviruses. PCR primers can be directed against variable regions of viral nucleic
acids, in order to ensure detection of a specific virus or viral strain, or against
conserved regions of viral sequences shared among strains or within a family, in
order to increase the opportunity for detecting multiple related viruses. Standard
PCR analysis can be coupled with hybridization methods to increase its versatility,
sensitivity and specificity. For example, the use of probes to various regions of the
amplicon might be useful for identifying the virus strain or family. However, PCR
methods have the limitation that viral genes may not be sufficiently conserved
among all members of a particular viral family for the genes to be detected even
when conserved regions are selected.
New and sensitive molecular methods with broad detection capabilities
are being developed. These are not yet in routine use but, as they become widely
available and validated, they will play an increasing role in the evaluation of cell
substrates. The sensitivity of these methods, as well as their breadth of detection,
should be considered when evaluating their applicability. One of the advantages
of some of these new methods is that they have the potential to discover new
viruses. These new approaches involve either degenerate PCR for whole virus
158
Annex 3
families or random-priming methods, which do not depend on a known sequence.

Analysis of the resulting amplicons has employed sequencing, hybridization to
oligonucleotide arrays, and mass spectrometry (103–105). The new generation of
massively parallel sequencing (MPS) methods may have particular utility. They
can be applied to detect virions following nuclease treatment to remove cellular
DNA and unencapsidated genomes. In this mode, MPS has been used to discover
new viruses in serum and other tissues and has revealed the contamination of
human vaccines by porcine circovirus (103, 106–110). MPS can also be employed
to screen cell substrates for both latent and lytic viruses by sequencing the
transcriptome. In this mode, enormous quantities of data are generated and
robust bioinformatic methods are required to detect viral sequences by either
positive selection against viral databases or negative selection to remove cellular
sequences (103, 110, 111). Care is required to exclude false “hits” to viruses due
to recognition of transduced cellular sequences present in some viral genomes,
or due to viral genes like virokines that have a close homology to cellular genes
(103, 105, 111).
It is probable that application of methods of this type will be expected
or required by regulatory agencies in future. At present the methods have
not been evaluated for sensitivity and specificity and should be thought of as
powerful investigational tools that can reveal issues that can be explored by more
established methods.
B.11.3 Bacteria, fungi, mollicutes and mycobacteria

The most common contaminants of cell culture are non-viral. These can be
introduced easily from the environment, materials, personnel, etc. Furthermore,
many such organisms multiply rapidly and can be pathogenic for humans. It
is also important in risk evaluation for the manufacturer to bear in mind that
standard compendial tests for “sterility” are intended to give an indication of
the effectiveness of aseptic processing in preventing general bacterial or fungal
contamination and are not capable of isolating all potential bacterial and fungal
contaminants. The manufacturer should consult with the NRA/NCL regarding
any particular materials or environments where there may be an elevated hazard
of contamination with particular types of fastidious organisms.
Biological starting materials, like cell substrates, should be characterized
to ensure that they are free of adventitious infectious organisms such as bacteria,
fungi, cultivable and non-cultivable mycoplasmas, spiroplasmas (in the case
of insect cells or cells exposed to plant-derived materials) and mycobacteria.
For a substance to be considered free of such contaminants, the assays should
demonstrate, at a predefined level of sensitivity, that a certain quantity of the
substance does not contain detectable levels of the contaminant. Testing should be
conducted in an aseptic environment under appropriate clean-room conditions,
to avoid false-positive results. Testing should include a plan to allow for repeat
159
testing to deal with potentially false-positive results and a prequalification plan

for reagents used in the tests.
Mycobacterial testing may be applied to cell-bank characterization if
the cells are susceptible to infection with Mycobacterium tuberculosis or other
species. Such testing should also be performed on primary cell cultures. It may
be necessary to lyse the host cells in order to detect mycobacteria, because some
strains may be primarily intracellular.
Detection of mycoplasma or spiroplasma may require different growth
conditions from methods used for mammalian cells, although at least one
– spiroplasma – can be cultivated at 30 °C. Positive controls for these tests
(particularly for spiroplasmas) are an issue that needs to be resolved. Spiroplasmas
have been reported as infectious agents in a number of insect species, and insect
cell lines have also been reported to cause pathogenic effects in mammals.
B.11.3.1 Bacterial and fungal sterility

Tests are performed as specified in Part A, section 5.2 of the Requirements for
biological substances no. 6 (112) by a method approved by the NRA/NCL.
Additional information can be found in national pharmacopoeias and ICH
documents (10, 113–115). For the MCB and WCB, the test is carried out using
for each medium 10 ml of supernatant fluid from cell cultures. In addition, the
test is carried out on at least 1% of the filled containers (i.e. cyropreservation vials)
with a minimum of two containers. For supernatant fluid, it is recommended
to use the membrane filtration method. For cell bank vial testing, it may be
necessary to use the direct inoculation method. Bacteriostasis and fungistasis
should be excluded.
Applicability

B.11.3.2 Mollicutes
Mollicutes are distinguished by an absence of a cell wall and include mycoplasmas,
acholeplasmas, spiroplasmas and others. They are parasites of various animals and
plants, living on or in the host’s cells. Mollicutes are also a frequent contaminant
of cell cultures. In addition to their potential pathogenicity, mycoplasmas
compete for nutrients, induce chromosomal abnormalities, interrupt metabolism
and inhibit cell fusion of host cells. M. pneumoniae is pathogenic for humans,
although there are no reported cases of human infections with this organism
arising from exposure to cell cultures or cell-derived products. In any case, cell
banks should be demonstrated to be free of such contamination, in order to be
suitable for the production of biologicals.
160
Annex 3
B.11.3.2.1 Mycoplasma and acholesplasma

Tests for mycoplasmas are performed as specified in Part A, sections 5.2 and 5.3 of
the Requirements for biological substances no. 6 (116), or by a method approved
by the NRA/NCL. Both the culture method and the indicator cell-culture
method should be used. NAT alone, in combination with cell culture or with an
appropriate detection method, may be used as an alternative to one or both of the
other methods, after suitable validation and discussion with the NRA/NCL. In this
case, a comparability study should be carried out. The comparability study should
include a comparison of the respective detection limits of the alternative method
and official methods. Specificity (mycoplasma panel detected, potential false-
positive results due to cross-reaction to other bacteria) should also be considered.
More details are available in the European pharmacopoeia Chapter 2.6.7 (117).
One or more containers of the MCB and each WCB are used for the test.
Applicability

B.11.3.2.2 Spiroplasma and others

Other mollicutes such as spiroplasma may be introduced into cell substrates
through contamination of raw materials (peptons) or due to the nature
and permissivity of the cells (e.g. insect cells). According to the cell bank
manufacturing process, if the raw material exposure is at the level of MCB or
before, it may be appropriate to test the MCB only. If further exposure is possible,
testing of the WCB may also be necessary.
Detection of such mollicutes may require adapted culture conditions
(medium and/or temperature), depending on the strain to be detected. To
guarantee a broad detection of the mollicutes, it is helpful to use NAT after
suitable validation with an appropriate model (e.g. Spiroplasma citri or other strain
according to the cell origin).
Applicability
■ Cell banks: MCB, WCB (recommended for insect cells)

■ Cell types: DCL, SCL, CCL (recommended for insect cell substrates
and when raw materials of plant origin are used during the cell bank
preparation or production process)
B.11.3.3 Mycobacteria
The test for mycobacteria is performed as described below or by a method
approved by the NRA/NCL.
161
Inoculate 0.2 ml of the sample in triplicate on to each of two suitable solid

media (such as Löwenstein–Jensen medium and Middlebrook 7H10 medium).
Inoculate 0.5 ml in triplicate into a suitable liquid medium at 37 °C for 56 days.
In some countries, the incubation period is 42 days.
An appropriate positive-control test should be conducted simultaneously with the

sample under evaluation, and the test should be shown to be capable of detecting
the growth of small amounts of mycobacteria. In addition, the fertility of the
medium in the presence of the preparation to be examined should be established
by a spiking inoculation of a suitable strain of a Mycobacterium sp., such as bacille
Calmette–Guérin (BCG). If at the end of the incubation time, no growth of
mycobacteria occurs in any of the test media, and the positive control and spiked
control show appropriate growth, the preparation complies with the test.
NAT may be used as an alternative to this culture method, provided that
such an assay is shown to be comparable to the compendial culture method. An
appropriate comparability study should be carried out that includes a comparison
of the respective detection limits of the alternative method and culture method.
Specificity (mycobacteria panel detected, potential false-positive results due to
cross-reaction to other bacteria) should also be considered. An in vivo method,
as described in the test in guinea-pigs, may also be used (see section B.11.2.1.3).
Applicability
■ Cell banks: MCB or WCB

B.11.4 Transmissible spongiform encephalopathies

TSEs are a group of slowly developing fatal neurological diseases affecting the
brains of animals and humans. The accepted view at present is that they are
caused by non-conventional infectious agents known as prions (PrP tse), which

are made up of a normal host protein (PrP) in an abnormal conformation. TSEs
include BSE of cattle, scrapie of sheep, CJD and its variant form (vCJD), GSS and
FFI in humans, CWD in elk and deer, and transmissible mink encephalopathy
(118, 119). Normal PrP (PrPc) protein may be expressed on cell surfaces, but
in vivo this protein can misfold and become the abnormal disease-causing type
PrP tse, which is able to catalyse the conversion of PrPc protein into the abnormal
conformation. Compared with PrPc, PrP tse is relatively resistant to common
proteolytic enzymes such as proteinase K.
BSE was first described in the United Kingdom in 1984, and the numbers
of clinical cases there reached a peak in 1992–1993. Other countries were also
affected. Currently, the number of new infections detected annually is low (120).
162
Annex 3
However, BSE remains a particular concern because cases still occur, albeit at a
low rate, and there is a legacy arising from the prolonged incubation period of the
disease, the life expectancy of cell banks, and the complexity of the processes by
which they are established.
BSE in cattle has been transmitted to humans in the form of vCJD.
Approximately 200 individuals have been affected either directly through
exposure to BSE-infected material or through secondary transmission by non-
leukocyte-depleted red blood cells. Classical CJD has also been transmitted by
medical procedures, including administration of cadaveric growth hormone
(121), corneal transplant and the use of dura mater, and vCJD may be transmissible
by the same routes. vCJD has also been transmitted by human blood products.
Although there is no evidence of vCJD transmission by plasma products,
public health precautions have been implemented to minimize the possible
risk of onward vCJD transmission by this route (122). Cattle-derived proteins,
including serum, have often been used in the growth of cells in culture and the
production of biological products, including vaccines and recombinant products.
Thus, it is important to ensure that any ruminant-derived material used in
biopharmaceutical manufacture is free of the agents that cause TSE. Moreover,
as there is a possible but unquantifiable risk that cells can become infected by
the agents of TSE, it is important that possibly contaminated ruminant material
should be excluded from the start of the development of any cell line used. When
there is insufficient traceability in the legacy of a cell line, a risk assessment
should be undertaken to aid decision-making about the suitability of the cell line
for the intended use. There is currently no practical validated test that can be
used for biological products or cell line testing for the agents of TSE other than
infection of susceptible species, where the experiments are very difficult because
of the length of the incubation. More usable tests such as protein misfolding
cyclic amplification (PMCA), which is analogous to PCR for nucleic acids, and
epitope protection assays (123) are under investigation, but their performance
characteristics when used to detect TSE agents in biological products or cell lines
have not been defined. Strategies for minimizing risk have therefore focused so far
on sourcing materials from countries believed to be at very low risk of infection
and on substituting animal-derived materials with non-animal-derived materials.
B.11.4.1 Infectivity categories of tissues

Ruminant tissues are categorized by WHO and other scientific bodies such as
the EMA into three categories (category A: high infectivity; category B: lower
infectivity; category C: no detectable infectivity) (57, 124). Category A includes
brain and category C includes materials such as testes and bile. Assays of
improved sensitivity have shown infectivity in tissues such as muscle that were
163
previously thought to be free of infectious agents, and the implication is that,

while certain tissues contain large amounts of infectivity, many other tissues may
contain low levels that are difficult to detect (57, 124–126).
B.11.4.2 Control measures, sourcing and traceability

Where effective alternatives to ruminant-derived material are available, they
should be used in cell culture and manufacturing procedures. Examples include:
cell culture medium free of animal material; polysorbate and magnesium stearate
of plant origin; enzymes, such as rennet, of microbial origin (used in lactose
production); and recombinant insulin and synthetic amino acids. It should
be noted that recombinant materials may themselves be exposed to animal
materials, so this potential should be considered when choosing recombinant
materials as alternatives. However, it is not always possible to use ingredients
that are free of animal materials, and raw materials of non-ruminant origin.
For example, fetal bovine serum may have to be used in the development of cell
lines or for fermentation. Under these circumstances, the raw materials should
be sourced from countries classified by the World Organisation for Animal
Health (OIE) as negligible BSE risk (geographical BSE-risk level I, or GBR I), as
classified by the European Food Safety Authority. Raw materials of category C
may be sourced from countries that are classified as controlled risk, provided
there is assurance that no cross-contamination with materials of category A or
B could have occurred during collection and processing (with the caveat that,
while they have undetectable levels of infectivity, it could conceivably be present).
Manufacturers should maintain records so that the finished product from any
batch is traceable to the origin of any ruminant ingredient used in its manufacture
that may pose a risk of exposure to a TSE agent, and each ruminant ingredient is
traceable to the finished product. This includes ingredients used to develop and
produce the MCB and WCB and, as far as is possible, traceability should be to the
derivation of the cell line itself. This traceability in both directions is important
for appropriate regulatory action if new scientific research indicates that there
is a risk of TSE infectivity in the materials used, or if the use of the products is
associated with vCJD. Category A and B ruminant materials originating from
BSE-enzootic countries should not be used in the production of biologicals
under any circumstances. Because new BSE cases continue to occur despite feed
bans, because suitable tests for TSE agents in raw materials are not available, and
because developments in scientific research indicate the presence of pathological
prions in materials of category C, the best approach to TSE safety is not to use
animal-derived protein. The next best approach is to source raw materials from
countries classified as free of BSE, bearing in mind that cases may be detected
in future.
164
Annex 3
B.11.4.3 Tests
No suitable screening tests are currently available for TSE agents in raw materials
of human/ruminant origin similar to serological or PCR assays for screening
for viral agents. Newer tests are being developed to screen for the presence of
TSE agents in blood (such as PMCA, epitope-protection assay and others). Such
tests, once validated, could eventually become suitable for the screening of raw
materials and cell banks.
Approximately 15% of human TSEs are associated with inherited
mutations in the PrP gene. These familial, but transmissible, TSEs are associated
with around 30 known pathogenic mutations or with insertions and deletions
in the octapeptide-repeat region of PrP (127). The PrP gene of new human cell
substrates should be sequenced to exclude the presence of these genetic changes.
B.12 Summary of tests for the evaluation and

characterization of animal cell substrates
This section provides an overview of tests that are recommended for the evaluation
and characterization of animal cell substrates proposed for use in the production
of biological products. Not all of the tests are appropriate for all animal cell
substrates, but each of them should be considered and a determination made as to
its applicability for a given cell substrate in the context of its use to manufacture a
specific product. In addition, the point(s) at which a test should be applied needs
to be rationalized. The overall testing strategy should provide assurance that risks
have been mitigated to reasonable levels for the product and for its intended use.
The testing strategy should be agreed with the NRA/NCL.
B.12.1 Cell seed

The cell seed is generally derived from a cell or tissue source of interest because
of its potential utility in the development of a biological product. In some cases,
the cell may be expected to serve as the substrate for the production of multiple
products. RCBs (see also section A.5.3) would be considered cell seeds. The cell
seed is usually of limited quantity, so extensive testing is not feasible. Some of
the seed is therefore used to produce a supply of cells in a quantity that allows
more extensive testing, as well as providing a long-term source of cells for use in
manufacturing. This secondary cell source is usually called the MCB. However,
the cell seed also may be used to produce additional low-passage material that
can be banked (pre-MCB) and used to generate MCBs that then are characterized
as described in this document.
Tests on the cell seed can be carried out at any point before the
establishment of the MCB. Usually, such tests are limited to obtaining information
that is essential for making the decision to commit resources to the preparation
of an MCB. Such tests typically include viability, morphology, identity (e.g.
165
karyotype, isoenzymes) and sterility (e.g. bacterial, fungal, mycoplasma). These

data serve as important background information, but they cannot substitute for
the full characterization of the MCB.
B.12.2 Master cell bank and working cell bank

In general the MCB will be developed to generate a sufficient quantity of cells
to supply enough vials of cells to produce many WCBs over an extended period
(usually years). MCBs typically contain at least 200 vials and often 1000 or more.
There should also be a sufficient number of vials in the WCB to provide material
for the characterization of the cell line.
Some tests on the WCB are conducted on cells recovered directly from
the bank itself; other tests are conducted on cells that have been propagated to
a passage at or beyond the level that will be used for production. In addition,
some tests may be appropriate to use as in-process control tests. In such cases,
they should be identified and described in the recommendations applicable to
specific products.
Authors
The scientific basis for the revision of the Requirements for the use of animal
cells as in vitro substrates for the production of biologicals published in WHO
Technical Report Series, No. 878, was developed at the meetings of the WHO Study
Group on Cell Substrates in 2006 and 2007 attended by the following people:
Dr P. Christian, National Institute for Biological Standards and Control,
Potters Bar, England; Dr M. Deschamps, GlaxoSmithKline Biologicals, Rixensart,
Belgium; Dr R.M. Dhere, Serum Institute of India, Pune, India; Dr C. Hutchens,
Pfizer, St. Louis, MO, USA; Dr J. Lebron, Merck Research Laboratories, West
Point, PA, USA; Dr I. Knezevic, World Health Organization, Geneva, Switzerland;
Dr S. Lambert, World Health Organization, Geneva, Switzerland; Dr A. Lewis,
Center for Biologics Evaluation and Research, Bethesda, MD, USA; Dr L. Mallet,
Sanofi Pasteur, Marcy L’Etoile, France; Dr P. Nandapalan, Therapeutic Goods
Administration, Woden, ACT, Australia; Dr K. Peden, Center for Biologics
Evaluation and Research, Bethesda, MD, USA; Dr J. Petricciani, Consultant,
Palm Springs, CA, USA; Dr R. Sheets, National Institute of Allergy and Infectious
Diseases, National Institutes of Health, Bethesda, MD, USA; Dr J. Shin, World
Health Organization, Geneva, Switzerland; Dr Y. Sohn, Korea Food and Drug
Administration, Seoul, Republic of Korea; Dr G. Stacey, National Institute for
Biological Standards and Control, Potters Bar, England; Mrs C.A.M. van der
Velden, Consultant, Groenekan, Netherlands; Dr R. Wagner, Paul-Ehrlich-Institute,
Langen, Germany; Dr O. Wimalaratne, Medical Research Institute, Colombo, Sri
Lanka; and Dr D.J. Wood, World Health Organization, Geneva, Switzerland.
166
Annex 3
Since then, several draft recommendations were prepared by the drafting

group consisting of: Dr J. Petricciani, Consultant, Palm Springs, CA, USA; Dr R.
Sheets, National Institute of Allergy and Infectious Diseases, National Institutes
of Health, Bethesda, MD, USA; Dr G. Stacey, National Institute for Biological
Standards and Control, Potters Bar, England; and Dr I. Knezevic, World Health
Organization, Geneva, Switzerland. These recommendations were reviewed by
the WHO Study Group on Cell Substrates in 2008.
Following the meeting of the Study Group on Cell Substrates in April
2009, draft recommendations were revised taking into account information on
the current manufacturing and regulatory practice provided at that meeting,
which was attended by the following participants:
Dr K.S. Ahn, Korea Food and Drug Administration, Seoul, Republic of
Korea; Dr J.H. Blusch, Novartis, Basel, Switzerland; Dr da Silva Guedes Jr., Bio-
Manguinhos/Fiocruz, Rio de Janeiro, Brazil; Dr M. Deschamps, GlaxoSmithKline
Biologicals, Wavre, Belgium; Dr G. Dong, National Institute for the Control of
Pharmaceutical and Biological Products, Beijing, China; Dr B. Gauvin, Amgen
Inc., Thousand Oaks, CA, USA; Dr H. Kavermann, Roche Diagnostics GmbH,
Penzberg, Germany; Dr K. King, United States Food and Drug Administration,
Bethesda, MD, USA; Dr I. Knezevic, World Health Organization, Geneva,
Switzerland; Dr A. Lewis, Center for Biologics Evaluation and Research, Bethesda,
MD, USA; Dr L. Mallet, Sanofi Pasteur, Marcy L'Etoile, France; Dr P. Minor,
National Institute for Biological Standards and Control, Potters Bar, England;
Dr P. Nandapalan, Therapeutic Goods Administration, Woden, ACT, Australia;
Dr D. Onions, Invitrogen Corporation, Carlsbad, CA, USA; Dr K. Peden, Center
for Biologics Evaluation and Research, Bethesda, MD, USA; Dr J. Petricciani,
Consultant, Palm Springs, CA, USA; Ms E. Ika Prawahju, National Agency of
Drug and Food Control, Jakarta, Indonesia; Dr R. Sheets, National Institute of
Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD,
USA; Dr G. Stacey, National Institute for Biological Standards and Control,
Potters Bar, England; Dr R. Wagner, Paul-Ehrlich-Institute, Langen, Germany;
and Dr D.J. Wood, World Health Organization, Geneva, Switzerland.
On the basis of the comments received from a broad range of regulators,
manufacturers of vaccines and other biologicals and other relevant experts in
2009, the draft recommendations were updated by the drafting group and posted
on the WHO biologicals web site for public consultation from 4 to 31 May 2010.
The WHO/BS/10.2132 document was prepared by the drafting group at
its meeting on 1–3 June 2010 at WHO, Geneva, taking into account comments
received from the reviewers and from the following meeting participants:
Dr C. Conrad, World Health Organization, Geneva, Switzerland; Dr H.
Kang, World Health Organization, Geneva, Switzerland; Dr I. Knezevic, World
Health Organization, Geneva, Switzerland; Dr J. Petricciani, Consultant, Palm
167
Springs, CA, USA; Dr R. Sheets, National Institute of Allergy and Infectious

Diseases, National Institutes of Health, Bethesda, MD, USA; Dr P. Minor,
National Institute for Biological Standards and Control, Potters Bar, England;
Dr D. Onions, Invitrogen Corporation, Carlsbad, CA, USA; Dr K. Peden, Center
for Biologics Evaluation and Research, Bethesda, MD, USA; Dr J. Shin, World
Health Organization, Geneva, Switzerland; and Dr Glyn Stacey, National Institute
for Biological Standards and Control, Potters Bar, England.
Further changes were made to document WHO/BS/10.2132 by the Expert
Committee on Biological Standardization, resulting in the present document.
Acknowledgements
Acknowledgements are due to the following experts for their useful comments
during several rounds of the draft recommendations:
Mr S.G. Bankar, Mr V.B. Vaidya, Serum Institute of India, Pune, India;
International Federation of Pharmaceutical Manufacturers and Associations
(Dr R. Field, Dr J. Kutza, Dr J. Liu, Dr D. Lindsay, AstraZeneca/MedImmune;
Dr K. Duffy, Dr P. Duncan, Dr L. Plitnick, Dr L. Vromans, Dr J. Wolf, Merck;
Dr R.W. Fedechko, Dr S. Pluschkell, Pfizer; Dr L. Scheppler, Crucell; Dr F.
Mortiaux, GlaxoSmithKline Biologicals; Dr B. Mouterde, Sanofi Pasteur; Dr R.
Krause, IFPMA); Dr J. Robertson, National Institute for Biological Standards
and Control, Potters Bar, England; Parenteral Drug Association (Dr A. Cundell,
Schering-Plough; Dr J-P. Gregersen, Novartis; Dr L. Hayflick, University of
California San Francisco; Dr L. Hendricks, Centocor; Dr A. Khan, Center for
Biologics Evaluation and Research; Dr R.W. Kozak, Bayer; Dr Z. Liu, Schering-
Plough; Dr B. Potts, Consultant; Dr M. Ruffing, Boehringer-Ingelheim; Dr S.
Seaver, Seaver Associates LLC; Dr D. Vacante, Centocor; Dr H. Willkommen,
Regulatory Affairs and Biological Safety Consulting; Dr M. Wisher, BioReliance;
Dr R. Wolff, Biologics Consulting Group); Dr K. Brusselmans, Dr G. Waeterloos,

Scientific Institute of Public Health, Brussels, Belgium.
References
1. Requirements for the use of animal cells as in vitro substrates for the production of biologicals.
In: WHO Expert Committee on Biological Standardization. Forty-seventh report. Geneva, World
Health Organization, 1998 (WHO Technical Report Series, No. 878), Annex 1.
2. Hilleman MR. Cells, vaccines, and the pursuit of precedent. National Cancer Institute Monograph,
1968, 29:463–470.
3. Hayflick L, Plotkin S, Stevenson R. History of the acceptance of human diploid cell strains as
substrates for human virus vaccine production. Developments in Biological Standardization, 1987,
68:9–17.
168
Annex 3
4. Requirements for poliomyelitis vaccine (inactivated). In: Requirements for biological substances:
1. general requirements for manufacturing establishments and control laboratories; 2. requirements
for poliomyelitis vaccine (inactivated). Report of a Study Group. Geneva, World Health Organization,
1959 (WHO Technical Report Series, No. 178).
5. Requirements for poliomyelitis vaccine (inactivated). In: Requirements for biological substances.
Report of a WHO Expert Group. Geneva, World Health Organization, 1966 (WHO Technical Report
Series, No. 323).
6. Hayflick L. History of the development of human diploid cell strains. In: Proceedings of the
Symposium on the Characterization and Uses of Human Diploid Cell Strains, Opatija, 1963,
Yugoslavia. Ljubljana, Casopisno podjetje, 1963:37–54.
7. Hayflick L. A brief history of cell substrates used for the preparation of human biologicals.
Developmental Biology, 2001, 106:5–23.
8. Suggested methods for the management and testing of diploid cell culture used for virus vaccine
production. In: Proceedings of the Symposium on Human Diploid Cells, Zagreb, 1970 Yugoslavia.
Zagreb, Izdavacki zavod Jugoslavenske akademije, 1970:213–216.
9. Requirements for poliomyelitis vaccine (oral). (requirements for biological substances, No. 7). In:
WHO Expert Committee on Biological Standardization. Twenty-fourth report. Geneva, World Health
Organization, 1972 (WHO Technical Report Series, No. 486), Annex 1.
10. Derivation and characterization of cell substrates used for production of biotechnological/
biological products. Geneva, International Conference on Harmonisation, 1997 (http://www.ich.
org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Quality/Q5D/Step4/Q5D_Guideline.
pdf, accessed 18 June 2011).
11. Guidance for industry. Characterization and qualification of cell substrates and other biological
starting materials used in the production of viral vaccines for the prevention of infectious diseases.
Rockville, MD, United States Food and Drug Administration, 2010 (http://www.fda.gov/
downloads/BiologicsBloodVaccines/GuidanceComplianceRegulatoryInformation/Guidances/
Vaccines/UCM202439.pdf, accessed 18 June 2011).
12. Acceptability of cell substrates for production of biologicals. Report of a WHO Study Group on
Biologicals. Geneva, World Health Organization, 1987 (WHO Technical Report Series, No. 747).
13. Requirements for continuous cell lines used for biologicals production. In: WHO Expert Committee
on Biological Standardization. Thirty-sixth report. Geneva, World Health Organization, 1987 (WHO
Technical Report Series, No. 745), Annex 3.
14. Viral safety evaluation of biotechnology products derived from cell lines of human or animal origin.
Geneva, International Conference on Harmonisation, 1999 (http://www.ich.org/fileadmin/Public_
Web_Site/ICH_Products/Guidelines/Quality/Q5A_R1/Step4/Q5A_R1__Guideline.pdf, accessed
18 June 2011).
15. Analysis of the expression construct in cells used for production of r-DNA derived protein products.
Geneva, International Conference on Harmonisation, 1995 (http://www.ich.org/fileadmin/Public_
Web_Site/ICH_Products/Guidelines/Quality/Q5B/Step4/Q5B_Guideline.pdf, accessed 18 June
2011).
16. Requirements for diphtheria, tetanus, pertussis and combined vaccines (revised 1989). In: WHO
Expert Committee on Biological Standardization. Fortieth report. Geneva, World Health Organization,
1990 (WHO Technical Report Series, No. 800), Annex 2.
17. Schaeffer WI. Terminology associated with cell, tissue, and organ culture, molecular biology,
and molecular genetics. Tissue Culture Association Terminology Committee. In Vitro Cellular and
Developmental Biology, 1990, 26:97–101.
169
18. Hendriksen CFM, ed. Laboratory animals in vaccine production and control: replacement, reduction
and refinement. Dordrecht, Kluwer Academic Publishers, 1988.
19. Jacobs JP. Updated results on the karyology of the WI-38, MRC-5 and MRC-9 cell strains.
Developments in Biological Standardization, 1976, 37:155–156.
20. Jacobs JP et al. Guidelines for the acceptability, management and testing of serially propagated
human diploid cells for the production of live virus vaccines for use in man. Journal of Biological
Standardization, 1981, 9:331–342.
21. Petricciani JC et al. Karyology standards for rhesus diploid cell line DBS-FRhL-2. Journal of
Biological Standardization, 1976, 4:43–49.
22. Schollmayer E et al. High resolution analysis and differential condensation in RBA-banded human
chromosomes. Human Genetics, 1981, 59:187–193.
23. Rønne M. Chromosome preparation and high resolution banding techniques: a review. Journal of
Dairy Science, 1989, 72:1363–1377.
24. Healy LE, Ludwig TE, Choo A. International banking: checks, deposits, and withdrawals. Cell Stem
Cell, 2008, 2(4):305–306.
25. Laflamme MA et al. Cardiomyocytes derived from human embryonic stem cells in pro-survival
factors enhance function of infarcted rat hearts. Nature Biotechnology, 2007, 25(9):1015–1024.
26. Fleckenstein B, Daniel MD, Hunt RD. Tumour inductions with DNA of oncogenic primate
herpesviruses. Nature, 1978, 274:57–59.
27. Petricciani JC, Regan PJ. Risk of neoplastic transformation from cellular DNA: calculations using the
oncogene model. Developments in Biological Standardization, 1986, 68:43–49.
28. Temin HM. Overview of biological effects of addition of DNA molecules to cells. Journal of Medical
Virology, 1990, 31:13–17.
29. Wierenga DE, Cogan J, Petriccaini JC. Administration of tumour cell chromatin to immunosuppressed
and non-immunosuppressed primates. Biologicals, 1995, 23:221–224.
30. Petricciani JC, Horaud FN. DNA, dragons, and sanity. Biologicals, 1995, 23:233–238.
31. Nichols WW et al. Potential DNA vaccine integration into host cell genome. Annals of the New York
Academy of Sciences, 1995, 772:30–38.
32. Kurth R. Risk potential of the chromosomal insertion of foreign DNA. Annals of the New York
Academy of Sciences, 1995, 772:140–150.
33. Coffin JM. Molecular mechanisms of nucleic acid integration. Journal of Medical Virology, 1990,
31:43–49.
34. Sheng L et al. Oncogenicity of DNA in vivo: tumour induction with expression plasmids for
activated H-ras and c-myc. Biologicals, 2008, 36:184–197.
35. Garnick RL. Experience with viral contamination in cell culture. Developments in Biological
36. Burstyn DG. Contamination of the CHO cells by orbiviruses. Developments in Biological
37. Rabenau H et al. Contamination of genetically engineered Chinese hamster ovary cells. Biologicals,
1993, 21(3):207–214.
38. Sheng-Fowler L, Lewis AM Jr, Peden K. Quantitative determination of the infectivity of the
proviral DNA of a retrovirus in vitro. Evaluation of methods for DNA inactivation. Biologicals, 2009,
37:259–269.
170
Annex 3
39. Lewis AM Jr., Krause P, Peden K. A defined-risks approach to the regulatory assessment of the
use of tumourigenic cells as substrates for viral vaccine manufacture. Developmental Biology,
2001, 106:513–535.
40. Krause PR, Lewis AM Jr. Safety of viral DNA in biological products. Biologicals, 1998, 26:317–320.
41. Sheng-Fowler L, Lewis AM Jr., Peden K. Issues associated with residual cell-substrate DNA in viral
vaccines. Biologicals, 2009, 37:190–195.
42. Ledwith BJ et al. Plasmid DNA vaccines: investigation of integration into host cellular DNA
following intramuscular injection in mice. Intervirology, 2000, 43:258–272.
43. Sheng-Fowler L et al. Tumors induced in mice by direct inoculation of plasmid DNA expressing
both activated H-ras and c-myc. International Journal of Biological Sciences, 2010, 6:151–162.
44. Burns PA et al. Transformation of mouse skin endothelial cells in vivo by direct application of
plasmid DNA encoding the human T24 H-ras oncogene. Oncogene, 1991, 6:1973–1978.
45. Meng F et al. MicroRNA-21 regulates expression of the PTEN tumour suppressor gene in human
hepatocellular cancer. Gastroenterology, 2007, 133:647–658.
46. Ma L, Teruya-Feldstein J, Weinberg RA. Tumour invasion and metastasis initiated by microRNA-
10b in breast cancer. Nature, 2007, 449:682–688.
47. He L et al. A microRNA polycistron as a potential human oncogene. Nature, 2005, 435:828–833.
48. Hayashita Y et al. A polycistronic microRNA cluster, miR-17-92, is over expressed in human lung
cancers and enhances cell proliferation. Cancer Research, 2005, 65:9628–9632.
49. Israel MA et al. Biological activity of polyoma viral DNA in mice and hamsters. Journal of Virology,
1979, 29:990–996.
50. Peden K et al. Biological activity of residual cell-substrate DNA. Developmental Biology (Basel),
2006, 123:45–53.
51. Perrin P, Morgeaux S. Inactivation of DNA by beta-propiolactone. Biologicals, 1995, 23:207–211.
52. Lebron JA et al. Adaptation of the WHO guideline for residual DNA in parenteral vaccines
produced on continuous cell lines to a limit for oral vaccines. Developmental Biology (Basel),
2006, 123:35–44.
53. Oh YK et al. Nasal absorption and biodistribution of plasmid DNA: an alternative route of DNA
vaccine delivery. Vaccine, 2001, 19:4519–4525.
54. Coecke S et al. Guidance on good cell culture practice. A report of the Second ECVAM Task Force
on Good Cell Culture Practice. Alternatives to Laboratory Animals/ATLA, 2005, 33:1–27.
55. Consensus guidance for banking and supply of human embryonic stem cell lines for research
purposes. Stem Cell Reviews and Reports, 2009, 5:301–314.
56. Stacey GN. Risk assessment of cell culture processes. In: Stacey GN, Davis JM, eds. Medicines from
animal cell culture. Chichester, J Wiley & Sons, 2007:567–587.
57. WHO guidelines on tissue infectivity distribution in transmissible spongiform encephalopathies.
Geneva, World Health Organization, 2006 (http://www.who.int/bloodproducts/tse/WHO%20
TSE%20Guidelines%20FINAL-22%20JuneupdatedNL.pdf, accessed 18 June 2011).
58. Plavsic ZM, Bolin S. Resistance of porcine circovirus to gamma irradiation. BioPharm International,
2001, 14:32–36.
59. Requirements for the collection, processing and quality control of blood, blood components and
plasma derivatives. In: WHO Expert Committee on Biological Standardization. Forty-third report.
Geneva, World Health Organization, 1994 (WHO Technical Report Series, No. 840), Annex 2.
171
60. Stacey GN, Masters JR. Cryopreservation and banking of mammalian cell lines. Nature Protocols,
2008, 3(12):1981–1989.
61. Chu L, Robinson DK. Industrial choices for protein production by large-scale cell culture. Current
Opinion in Biotechnology, 2001, 12:180–187.
62. Ho Y, Varley J, Mantalaris A. Development and analysis of a mathematical model for antibody-
producing GS-NSO cells under normal and hyperosmotic culture conditions. Biotechnology
Progress, 2006, 22:1560–1569.
63. McPherson I, Montagnier I. Agar suspension culture for the selective assay of cells transformed by
polyoma virus. Virology, 1964, 23:291–294.
64. Petricciani JC, Levenbook I, Locke R. Human muscle: a model for the study of human neoplasia.
Investigational New Drugs, 1983, 1:297–302.
65. Giovanella BC et al. Development of invasive tumours in nude mice after injection of cultured
human melanoma cell. Journal of the National Cancer Institute, 1972, 48:1531–1534.
66. Manohar M et al. Assessing the tumorigenic phenotype of VERO cells in adult and newborn nude
mice. Biologicals, 2008, 36:65–72.
67. Hentze H et al. Teratoma formation by human embryonic stem cells: evaluation of essential
parameters for future safety studies. Stem Cell Research, 2009, 2:198–210.
68. van Steenis G, van Wezel AL. Use of ATG-treated newborn rat for in vivo tumorigenicity testing of
cell substrates. Developments in Biological Standardization, 1982, 50:37–46.
69. Noguchi PD, Johnson JB, Petricciani JC. Comparison of the nude mouse and immunosuppressed
newborn hamsters as a quantitative tumorigenicity test for human cell lines. Pathology, 1979,
7:302–304.
70. Wallace R, Vasington PJ, Petricciani JC. Heterotransplantation of cultured cell lines in newborn
hamsters treated with antilymphocyte serum. Nature, 1971, 230:454–455.
71. Foley GE et al. Assessment of potential malignancy of cultured cells: further observations on the
differentiation of “normal” and “neoplastic” cells maintained in vitro by heterotransplantation in
Syrian hamsters. National Cancer Institute Monograph, 1962, 7:173–204.
72. Petricciani JC, Martin DP. Use of nonhuman primates for assaying tumorigenicity of viral-vaccine
cell substrates. Transplantation Proceedings, 1974, 6:189–192.
73. Furesz J et al. Tumorigenicity testing of various cell substrates for production of biologicals.
Developments in Biological Standardization, 1989, 70:233–243.
74. Levenbook I et al. Tumorigenicity testing of primate cell lines in nude mice, muscle organ culture
and for colony formation in soft agarose. Journal of Biological Standardization, 1985, 13:135–141.
75. Levenbook I, Smith PL, Petricciani JC. A comparison of three routes of inoculation for testing
tumorigenicity of cell lines in nude mice. Journal of Biological Standardization, 1981, 9:75–80.
76. Lewis AM Jr et al. Evaluating virus-transformed cell tumorigenicity. Journal of Virological Methods,
1999, 79:41–50.
77. Pulciani S et al. Transforming genes in human tumours. Journal of Cellular Biochemistry, 1982,
20:51–61.
78. Murray MJ et al. Three different human tumour cell lines contain different oncogenes. Cell, 1981,
25:355–361.
79. Shih C et al. Transforming genes of carcinomas and neuroblastomas introduced into mouse
fibroblasts. Nature, 1981, 290:261–264.
80. Wood DJ, Minor PD. Meeting report. Use of human diploid cells in vaccine production. Biologicals,
1990, 18:143–146.
172
Annex 3
81. Shaffer LG, Tommerup N, eds. An international system for human cytogenetic nomenclature. Basel,
Karger, 2005.
82. Recommendations for the production and control of poliomyelitis vaccine (oral). In: WHO Expert
Committee on Biological Standardization. Fiftieth report. Geneva, World Health Organization, 2002
(WHO Technical Report Series, No. 904), Annex 1.
83. Requirements for measles, mumps and rubella vaccines and combined vaccine (live) (requirements
for Biological Substances No. 47). In: WHO Expert Committee on Biological Standardization. Forty-
fourth report. Geneva, World Health Organization, 1994 (WHO Technical Report Series, No. 848),
Note.
84. Nicklas W, Kraft V, Meyer B. Contamination of transplantable tumours, cell lines, and monoclonal
antibodies with rodent viruses. Laboratory Animal Science, 1993, 43(4):296–300.
85. Collins MJ, Parker JC. Murine virus contaminants of leukemia viruses and transplantable tumours.
Journal of the National Cancer Institute, 1972, 49(4):1139–1143.
86. Bierley ST et al. A comparison of methods for the estimation of retroviral burden. Developments in
Biological Standardization, 1996, 88:163–165.
87. Goff SP. Retroviridae: the retroviruses and their replication. In: Knipe DM et al., eds. Fields virology,
5th ed. Baltimore, MD, Lippincott, Williams & Wilkins, 2007:199–2069.
88. Arnold BA, Hepler RW, Keller PM. One-step fluorescent probe product-enhanced reverse
transcriptase assay. Biotechniques, 1998, 25:98–106.
89. Maudru T, Peden KW. Adaptation of the fluorogenic 5'-nuclease chemistry to a PCR-based
reverse transcriptase assay. Biotechniques, 1998, 25:972–975.
90. Lovatt A et al. High throughput detection of retrovirus-associated reverse transcriptase using
an improved fluorescent product enhanced reverse transcriptase assay and its comparison to
conventional detection methods. Journal of Virological Methods, 1999, 82:185–200.
91. Sears JF, Khan AS. Single-tube fluorescent product-enhanced reverse transcriptase assay with
Ampliwax (STF-PERT) for retrovirus quantitation. Journal of Virological Methods, 2003, 108:139–
142.
92. Pyra H, Böni J, Schüpbach J. Ultrasensitive retrovirus detection by a reverse transcriptase assay
based on product enhancement. Proceedings of the National Academy of Sciences of the United
States of America, 1994, 91:1544–1548.
93. Böni J, Pyra H, Schüpbach J. Sensitive detection and quantification of particle-associated
reverse transcriptase in plasma of HIV-1-infected individuals by the product-enhanced reverse
transcriptase (PERT) assay. Journal of Medical Virology, 1996, 49:23–28.
94. Böni J et al. Detection of reverse transcriptase activity in live attenuated virus vaccines. Clinical
and Diagnostic Virology, 1996, 5:43–53.
95. Weissmahr RN, Schüpbach J, Böni J. Reverse transcriptase activity in chicken embryo fibroblast
culture supernatants is associated with particles containing endogenous avian retrovirus EAV-0
RNA. Journal of Virology, 1997, 71:3005–3012.
96. Khan AS et al. The reverse transcriptase activity in cell-free medium of chicken embryo fibroblast
cultures is not associated with a replication-competent retrovirus. Journal of Clinical Virology,
1998, 11:7–18.
97. Maudru T, Peden K. Elimination of background signals in a modified polymerase chain reaction-
based reverse transcriptase assay. Journal of Virological Methods, 1997, 66:247–261.
98. Lugert R et al. Specific suppression of false-positive signals in the product-enhanced reverse
transcriptase assay. Biotechniques, 1996, 20:210–217.
173
99. Voisset C et al. Specific detection of RT activity in culture supernatants of retrovirus-producing

cells, using synthetic DNA as competitor in polymerase enhanced reverse transcriptase assay.
Journal of Virological Methods, 2001, 94:187–193.
100. Fan XY et al. A modified single-tube one-step product-enhanced reverse transcriptase (mSTOS-
PERT) assay with heparin as DNA polymerase inhibitor for specific detection of RTase activity.
Journal of Clinical Virology, 2006, 37:305–312.
101. Peebles PT. An in vitro focus-induction assay for xenotropic murine leukemia virus, feline leukemia
virus C, and the feline-primate viruses RD-114/CCC/M-7. Virology, 1975, 67:288–291.
102. Sommerfelt MA, Weiss RA. Receptor interference groups of 20 retroviruses plating on human
cells. Virology, 1990, 176:58–69.
103. Onions D, Kolman J. Massively parallel sequencing, a new method for detecting adventitious
agents. Biologicals, 2010, 38:377–380.
104. Wang D et al. Viral discovery and sequence recovery using DNA microarrays. PLoS Biology, 2003,
1:E2.
105. Ecker DJ et al. Ibis T5000: a universal biosensor approach for microbiology. Nature Reviews,
Microbiology, 2008, 6:553–558.
106. Allander T et al. Cloning of a human parvovirus by molecular screening of respiratory tract
samples. Proceedings of the National Academy of Sciences of the United States of America, 2005,
102(43):12891–12896.
107. Gaynor AM et al. Identification of a novel polyomavirus from patients with acute respiratory
tract infections. PLoS Pathogens, 2007, 3(5):E64.
108. Allander T et al. A virus discovery method incorporating DNase treatment and its application
to the identification of two bovine parvovirus species. Proceedings of the National Academy of
Sciences of the United States of America, 2001, 98(20):11609–11614.
109. Allander T et al. Identification of a third human polyomavirus. Journal of Virology, 2007,
81(8):4130–4136.
110. Victoria JG et al. Viral nucleic acids in live-attenuated vaccines: detection of minority variants
and an adventitious virus. Journal of Virology, 2010, 84(12):6033–6040 (Epub 7 April 2010).
111. Sampath R et al. Global surveillance of emerging influenza virus genotypes by mass spectrometry.
PLoS One, 2007, 2:e489.
112. General requirements for the sterility of biological substances (requirements for biological
substances, No. 6, revised 1973). In: WHO Expert Committee on Biological Standardization. Twenty-
fifth report. Geneva, World Health Organization, 1973 (WHO Technical Report Series, No. 530),
Annex 4.
113. Sterility test. In: Code of Federal Regulations Title 21. Washington. DC, National Archives and
Records Administration, 2010, Section 610.12.
114. USP 35-NF 30, Sterility tests <71> . In: United States pharmacopeia. Rockville, MD, US
Pharmacopeial Convention, 2012:69–74.
115. Sterility. In: European pharmacopoeia, 6th ed. Strasbourg, EDQM Publications, 2007, Section 2.6.1.
116. General requirements for the sterility of biological substances (requirements for biological
substances, No. 6, revised 1973). In: WHO Expert Committee on Biological Standardization. Forty-
sixth report. Geneva, World Health Organization, 1998 (WHO Technical Report Series, No. 872),
Annex 3.
117. Mycoplasmas. In: European pharmacopoeia, 6th ed. Strasbourg, EDQM Publications, 2007, Section
2.6.7.
174
Annex 3
118. Kovacs GG, Budka H. Prion diseases: from protein to cell pathology. American Journal of Pathology,
2008, 172:555–565.
119. Ryou C. Prions and prion diseases: fundamentals and mechanistic details. Journal of Microbiology
and Biotechnology, 2007, 17:1059–1070.
120. Number of reported cases of bovine spongiform encephalopathy (BSE) in farmed cattle worldwide
(excluding the United Kingdom). Paris, World Organisation for Animal Health (http://www.oie.
int/en/animal-health-in-the-world/bse-specific-data/number-of-reported-cases-worldwide-
excluding-the-united-kingdom/, accessed 29 June 2011).
121. Preece M.A. Creutzfeldt-Jakob disease following treatment with human pituitary hormones.
Clinical Endocrinology, 1991, 34:527–529.
122. Millar CM et al. Risk reduction strategies for variant Creutzfeldt–Jakob disease transmission by UK
plasma products and their impact on patients with inherited bleeding disorders. Haemophilia,
2010, 16(2):305–315.
123. Minor P, Brown P. Diagnostic tests for antemortem screening of Creutzfeldt–Jakob disease. In:
Turner ML, ed. Creutzfeldt–Jakob disease: managing the risk of transmission by blood, plasma, and
tissues. Bethesda, MD, AABB, 2006:119–148.
124. WHO tables on guidelines on tissue infectivity distribution in transmissible spongiform
encephalopathies. Geneva, World Health Organization, 2010 (WHO/EMP/QSM/2010.1; http://
www.who.int/bloodproducts/tablestissueinfectivity.pdf, accessed 18 June 2011).
125. Update of the opinion on TSE infectivity distribution in ruminant tissues. Brussels, Health and
Consumer Protection Directorate-General, European Commission, 2002 (http://ec.europa.eu/
food/fs/sc/ssc/out296_en.pdf, accessed 18 July 2011).
126. Note for guidance on minimising the risk of transmitting animal spongiform encephalopathy
agents via human and veterinary medicinal products (EMA/410/01 rev.3). Official Journal of
the European Union, 5 March 2011, C 73/1–C 73 18 (http://www.emea.europa.eu/docs/en_GB/
document_library/Scientific_guideline/2009/09/WC500003700.pdf, accessed 28 February 2013).
127. McKintosh E, Tabrizi SJ, Collinge J. Prion diseases. Journal of Neurovirology, 2003, 9(2):183–193.
175
Appendix 1
Tests for bovine viruses in serum used to produce cell
banks
Serum should be tested for adventitious agents such as bacteria, fungi,
mycoplasmas and viruses, prior to use in the production of MCBs and WCBs.
In addition, consideration should be given to risk-mitigation strategies, such as
inactivation by heat or irradiation, to ensure that adventitious agents that were not
detected in the manufacture and quality control of the serum will be inactivated
to a degree acceptable to the NRA/NCL. If irradiation or other inactivation (e.g.
heat sterilization) methods are used in the manufacture of the serum, the tests
for adventitious agents should be performed prior to inactivation, to enhance the
opportunity for detecting the contamination. If evidence of viral contamination
is found in any of the tests, the serum is acceptable only if the virus is identified
and shown to be present in an amount that has been shown in a validation study
to be effectively inactivated. For serum that is not to be subjected to a virus
inactivation/removal procedure, if evidence of viral contamination is found in
any tests, generally the serum will not be acceptable. If the manufacturer chooses
to use serum that has not been inactivated, thorough testing of the serum for
adventitious agents, using current best practices, should be undertaken. If any
viruses are identified in the serum, the cell banks made in this manner should be
shown to be free of the identified virus(es).
If irradiation is used, it is important to ensure that a reproducible dose
is delivered to all batches and to the component units of each batch.
The irradiation dose must be low enough for the biological properties
of the reagents to be retained, while being high enough to reduce

virological risk. Therefore, irradiation delivered at such a dose may not
be a sterilizing dose.
Factors to be considered in testing serum

Bovine serum can be contaminated by a wide range of viruses. Manufacturers
typically produce very large pools of serum involving samples from up to a
thousand animals. Consequently, many serum batches contain detectable
genomic sequences of viruses such as BVDV and bovine polyomavirus (1),
although this may represent contamination of the pool by one viraemic animal.
Other viruses are sporadic contaminants and may be regionally restricted,
such as Cache Valley virus, BTV and epizootic haemorrhagic disease virus. In
176
Annex 3
some cases, contamination has been reported only on a few occasions, as in the
case of calicivirus 2117 (2).
Application of new methods such as MPS has revealed new viruses,
like the parvoviruses, some of which are frequent and high-level contaminants
of serum (3, 4). The importance and potential pathogenicity of these viruses
requires further investigation.
An important factor in infectivity assays is that virions might be
neutralized by antibody in the serum pool. It is advisable to set limits for the level
of BVDV-neutralizing antibody in serum pools, as this may mask the presence
of potentially infectious virus.
There should be awareness of the statistical limits of screening assays
in detecting viruses in large serum pools. For example, in an infection of a
fermenter by Cache Valley virus, it was estimated that fewer than 10 viruses per
litre were present in the serum and, at this low level, the virus escaped detection
by conventional screening methods (5).
General screening assay for infectious viruses

A general screening assay typically involves culturing indicator cells over 21 days
with test serum at 15% in the medium. At least two subpassages of the cells should
be undertaken, usually at days 7 and 14. Detection of virus infection involves
regular examination for the development of a CPE, haemadsorption assays and
immunofluorescence (or other appropriate immunological detection method) for
specific viruses. Immunofluorescence is particularly important for the detection
of BVDV, as many isolates are non-cytopathic. At the end of the assay, cytological
staining (e.g. with Giemsa stain) is used to reveal viral inclusions and other CPEs
that were not detected during the direct observation of the live cells.
Indicator cells should be selected that are permissive for a wide range of
bovine viruses. MDCK cells or bovine turbinate cells are often used, and it is also
of value to include additional cells such as Vero cells.
The assay should be capable of detecting the following: BVDV, BPIV3,
BPV1, rabies virus, REO3, IBR, BRSV, BTV, bovine adenovirus 5 (BAV5), and
vesicular stomatitis virus. Separate positive-control bottles of indicator cells
should be infected with each of the viruses above, except rabies virus. In the case
of rabies virus, slides of fixed infected cells should be used as a positive control
for the immunofluorescent assay. Uninfected negative-control cells should also
be established.
A typical assay involves the use of 75 cm2 bottles containing the indicator
cells and a total of ~250 ml of test serum, allowing for serum used during refeeding
of the cells after passage.
177
Procedure
Assay set-up
Initially, negative-control bottles and test article bottles are established. The test
article bottles are inoculated and maintained with the test serum at 15% in the
medium. The negative-control bottles are mock infected with serum known to be
free of detectable viruses. Passage of the cells is usually required on day 7.
Cells for the positive control are prepared from the negative control
bottles on day 13 or 14 or when the cells are ≥70% confluent. The cells are
subcultured into 25 cm2 flasks (for immunofluorescence) and six-well plates (for
haemadsorption and cytological staining).
The following day, the remaining negative-control and test article cells
are subcultured to 75 cm2 flasks for immunofluorescence and to six-well plates
for haemadsorption and cytological staining.
Infection with positive controls

Coincident with the final subculture of test article and negative-control cultures,
flasks of bovine turbinate cells are inoculated with the immunofluorescence
positive control viruses BVDV, BAV5, BPV, BTV, BRSV, IBR and BPIV3. Plates
of bovine turbinate cells are inoculated with BPIV3, the positive control for
haemadsorption, and with cytopathic BVDV, the positive control for cytological
staining. Likewise, Vero flasks are inoculated with REO3, the immunofluorescence
positive control, and plates are inoculated with BPIV3, the haemadsorption and
cytological staining positive control. All immunofluorescence positive-control
viruses should be inoculated at 100–300 TCID50 (median tissue culture infective
dose).
Analysis
After a minimum of 21 days after inoculation, and at least 7 days after the last
subculture (but earlier if CPE is observed), negative-control and test article cultures
are assayed for haemadsorption and fixed for immunofluorescence and cytological
staining. Cells from the positive-control flasks are transferred to multiwell slides
and fixed for immunofluorescence when CPE involving ≥10% of the monolayer
is observed, and stored at ≤–60 °C. Cells in the positive-control six-well plates are
assayed for haemadsorption and cytological staining 7 days after inoculation, or
when CPE is apparent. Haemadsorption involves testing at least one six-well plate
with chicken and guinea-pig erythrocytes at 2–8 °C and at 20–25 °C.
Nucleic acid amplification assays for viruses

Nucleic acid amplification technologies such as PCR have utility in screening
serum for sporadic contaminants and for those viruses where infectivity assays
are not available. Nucleic acid extractions should be from a significant volume
178
Annex 3
(e.g. 25–50 ml) and the statistical limits for detection in the serum pool should be
calculated. The presence of genomic sequences does not necessarily indicate the
presence of infectious virus, although encapsidated genomes can be identified
by treatment of the sample with nucleases prior to amplification. Some virus-
inactivating or removal processes can be evaluated using NAT, by determining
whether intact, full-length, amplifiable genomes are present before and after
treatment.
Specific in vitro infectivity assays

Bovine polyomavirus is an important contaminant because it is able to infect
primate cells (6), belongs to an oncogenic family of viruses and expresses a
T-antigen that can transform primary cells into tumour cells (7). Furthermore,
there is serological evidence of zoonotic infection (8). Infectious virus is not
easily detected in conventional assays; a long period of culture and a NAT end-
point or immunological end-point such as immunofluorescence should be used.
Other viruses are not easily detected in standard infectivity assays. For
instance, calicivirus 2117 appears to be more permissive for replication in CHO
cells than standard bovine cell lines used in in vitro infectivity assays. Similarly,
while general screening methods will detect certain bovine adenoviruses,
herpesviruses and parvoviruses, not all bovine viruses belonging to these families
are detected.
References
1. Schuurman R et al. Frequent detection of bovine polyomavirus in commercial batches of calf
serum by using the polymerase chain reaction. Journal of General Virology, 1991, 72(11):2739–
2745.
2. Oehmig A et al. Identification of a calicivirus isolate of unknown origin. Journal of General Virology,
2003, 84(10):2837–2845.
3. Allander T et al. A virus discovery method incorporating DNase treatment and its application
to the identification of two bovine parvovirus species. Proceedings of the National Academy of
Sciences of the United States of America, 2001, 98(20):11609–11614 (Epub 18 September 2001).
4. Onions D, Kolman J. Massively parallel sequencing, a new method for detecting adventitious
agents. Biologicals, 2010, 38(3):377–380.
5. Nims RW et al. Detection of Cache Valley virus in biologics manufactured in CHO cells. BioPharm
International, 2008, 21(10).
6. Richmond JE, Parry JV, Gardner SD. Characterization of a polyomavirus in two foetal rhesus
monkey kidney cell lines used for the growth of hepatitis A virus. Archives of Virology, 1984, 80(2–
3):131–146.
7. Schuurman R et al. Bovine polyomavirus, a cell-transforming virus with tumorigenic potential.
Journal of General Virology, 1992, 73(11):2871–2878.
8. Parry JV, Gardner SD. Human exposure to bovine polyomavirus: a zoonosis? Archives of Virology,
1986, 87(3–4):287–296.
179
Appendix 2
Tumorigenicity protocol using athymic nude mice to
assess mammalian cells
During the characterization of an MCB (or WCB), the cells should be examined
for tumorigenicity in a test approved by the NRA or the NCL.
The following model protocol is provided to assist manufacturers and
NRAs/NCLs to standardize the tumorigenicity testing procedure so that the
interpretation and comparability of data between various laboratories and
regulatory authorities can be facilitated.
1. Test animals
The test article cell line and the control cells are each injected into separate
groups of 10 athymic mice (Nu/Nu genotype) 4–7 weeks old.
Because male athymic mice often display aggressive traits against each
other when housed together, loss of some mice during the observation
period often occurs. Therefore, the use of only female mice should be
considered.
2. Test article cells

Cells from the MCB or WCB that have been propagated to at least three population
doublings beyond the limit for production are examined for tumorigenicity.
3. Control cells
a. Positive control cells
HeLa cells from the WHO cell bank are recommended as the positive control
reference preparation. Portions of that bank are stored at the American Type
Culture Collection (USA) and the National Institute for Biological Standards and
Control (England).
Other cells may be acceptable to the NRA/NCL if HeLa cells from the
WHO cell bank are not available.
b. Negative-control cells
Negative-control cells are not required. Databases (both published data and the
unpublished records/data of the animal production facility that supplied the test
animals) of rates of spontaneous neoplastic diseases in nude mice may be taken
into account during the assessment of the results of a tumorigenicity test.
180
Annex 3
If negative-control cells are included, clear justification must be provided.

In particular, the number of animals used must provide meaningful data, and the
rationale for generating additional data must be persuasive to the NRA/NCL in
the context of animal welfare regulations.
4. Validity
In a valid test, progressively growing tumours should be produced in at least 9
out of 10 animals injected with the positive control reference cells. At least 90%
of the inoculated control and cells and test cells must be viable for the test to
be valid.
5. Inoculum
The inoculum for each animal is 10 7 viable cells (except as described in 11.b,
below), suspended in a volume of 0.1 ml PBS.
Cell culture medium without serum has been used in the past to suspend
the cell inoculum. However, many current media are serum free and
contain one or more growth factors that may affect the result of the
tumorigenicity assay. Therefore, careful consideration should be given to
the choice of the liquid into which the cells are suspended.
6. Injection route and site

The injection of cells may be by either the intramuscular or the subcutaneous
route. If the intramuscular route is selected, the cells should be injected into the
thigh of one leg. If the subcutaneous route is selected, the cells should be injected
into the supraclavicular region of the trunk.
On the basis of findings of published studies, the intracerebral route
may be more appropriate in some cases. For example, lymphoblastoid
cells have been shown to proliferate best when inoculated by the
intracerebral route.
7. Observation period
All animals are examined weekly by observation and palpation, for a minimum of
16 weeks (i.e. 4 months) for evidence of nodule formation at the site of injection
when the route of inoculation is intramuscular or subcutaneous. Examinations
need not be more frequent than two times a week for the first 3–6 weeks, and
once a week thereafter.
In some countries, the observation period is 4–7 months, depending
on the level of concern associated with the specific cell substrate in the
context of the product being developed. Whether a longer observation
period is needed should be agreed with the NRA/NCL. Also see 8 below.
181
8. Assessment of the inoculation site over time

If a nodule appears, it is measured in two perpendicular dimensions, the
measurements being recorded weekly to determine whether the nodule grows
progressively, remains stable or decreases in size over time. Animals bearing
nodules that appear to be regressing should not be killed until the end of the
observation period. Cell lines that produce nodules that fail to grow progressively
are not considered to be tumorigenic.
If a nodule fails to grow progressively but persists during the observation
period and retains the histopathological morphology of a neoplasm,
this should be discussed with the NRA/NCL, to determine whether
additional testing will be required. Such testing could include extending
the observation period or switching to a newborn nude mouse, ATS-
treated newborn rat, or other in vivo model, to assess the tumorigenicity
of the cell substrate.
If the cells that are injected fail to form tumours or to persist during
the 4-month observation period, it may be necessary to extend the
observation period or switch to a newborn nude mouse, ATS-treated
newborn rat, or other in vivo model to assess the tumorigenicity of the
cell substrate. This will depend on the level of concern associated with
the specific cell substrate in the context of the product being developed.
Whether such additional testing is needed should be agreed with the
NRA/NCL.
9. Final assessment of the inoculation site and other sites

At the end of the observation period, or at an earlier time if required due to the
death of an animal or other justifiable circumstances, all animals (including the
reference group(s)) are killed and examined for gross and microscopic evidence
of the proliferation of inoculated cells at the site of injection and at other sites
such as the heart, lungs, liver, spleen, kidneys, brain and regional lymph nodes,
since some CCLs may give rise to tumours at distant sites without evidence
of tumour at the injection site. The tissues are fixed in 10% formol saline and
sections are stained with heamatoxylin and eosin for histological examination to
determine whether there is evidence of tumour formation and metastases by the
inoculated cells.
10. Assessment of metastases (if any)

Any metastatic lesions are examined further to establish their relationship to
the primary tumour. If what appears to be a metastasis to a distant site differs
histopathologically from the primary tumour, consideration should be given to
the possibility that the tumour either developed spontaneously or was induced by
one or more of the components of the cell substrate, such as an oncogenic virus.
182
Annex 3
If the histopathology or genotype of any tumours that develop are

inconsistent with the inoculated cell type, or are of a histopathological
type that has not been recognized as occurring spontaneously in the test
species, additional tests should be undertaken to determine whether
such tumours are actually spontaneous or are induced by elements
within the cell substrate itself, such as oncogenic viruses or oncogenic
DNA sequences. In such cases, appropriate follow-up studies should be
discussed and agreed with the NRA/NCL.
11. Interpretation of results

a. The test in nude mice is considered positive if at least 2 out of 10 animals
inoculated with the test article cells develop tumours that meet the following
two criteria:
i. Tumours appear at the site of inoculation or at a metastatic site.
ii. Histological or genotypic examination reveals that the nature of
the cells constituting the tumours is consistent with that of the
inoculated cells.
In the past, chromosomal markers have been useful to demonstrate that
the tumour cells are of the same species as that from which the inoculated
cells were derived. However, the use of cytogenetics for this purpose has
largely been replaced by genetic and antigenic markers.
b. If only 1 out of 10 animals develops a tumour that meets the two criteria
in 11.a, the cell line should be considered to be possibly tumorigenic and
should be examined further. Such testing could include one or more of
the following: repeating the test in an additional 10 nude mice, extending
the observation period, increasing the size of the inoculum, or switching
to the newborn nude mouse model, the ATS-treated newborn rat model,
or other in vivo model. In such cases, appropriate follow-up investigations
should be discussed and agreed with the NRA/NCL. For example, it may be
appropriate to determine whether the tumour is of nude mouse origin and
whether there are any viral or inoculated cell DNA sequences present.
Assessment of dose–response may provide additional information on
the characteristics of the CCL. If such studies are undertaken, the design
should be based on the in vivo titration of the inoculum in groups of
10 animals per dose level. For example, if 10 out of 10 animals develop
tumours with an inoculum of 10 7 cells, the titration could be done with
10 5, 10 3 and 10 1 cells in groups of 10 animals each.
183
Appendix 3
Oncogenicity protocol for the evaluation of cellular DNA
and cell lysates
When appropriate, and particularly for vaccines, cell DNA and cell lysates
from tumorigenic cell substrates should be examined for oncogenicity in a test
approved by the NRA/NCL.
In some countries, the following testing strategy is used:
1. Type of test animals

Newborn (i.e. <3 days old) nude mice, newborn hamsters and newborn rats
have been used to assess the oncogenic potential of cell lines. At this stage, it
is not possible to draw definitive conclusions on the relative sensitivity of the
three animal assays for oncogenicity, and testing is recommended in each of
them. When data on the ability of these models to detect oncogenic activity are
obtained, this recommendation may change.
2. The point in cells’ life history at which they should be tested

Cells from the MCB or WCB, propagated to the proposed in vitro cell age
for production or beyond, should be examined for oncogenicity. Three extra
population doublings ensure that the results of the oncogenicity test can be used
in the assessment of overall safety of the product, even under the assumption of a
worst-case situation, and therefore provide a safety buffer.
3. Use of controls
The purpose of the positive control is to assure that an individual test is valid,
by demonstrating that the animal model has the capacity to develop tumours
from inoculated cell components (i.e. a negative result is unlikely to be due to a
problem with the in vivo test). While an appropriate positive control for cell lysate
oncogenicity assay is not clear, the recent description of an oncogene-expression
plasmid for activated H-ras and c-myc has been shown to induce tumours in
animals (1). As the test with cell lysates is designed primarily to detect oncogenic
viruses rather than oncogenes, the use of DNA as a positive control may not be
suitable, both because of the nature of the assay and because DNA may not be
stable in a cellular lysate.
Whether a negative-control arm, such as PBS, is included should be
discussed with the NRA/NCL. An advantage of including a negative-control arm
184
Annex 3
is that the frequency of tumour induction with lysates is expected to be low and
may approximate to the spontaneous tumour frequency in the indicator rodent,
providing an important comparison to the test article arm.
4. Number of test animals

While the number of animals in a tumorigenicity test can be 10 per group, the
number in an oncogenicity test should be larger, owing to the lower expected
tumour incidence. The number per group should be discussed with the NRA/
NCL.
5. Inoculation of test material

a. Cell lysate
A lysate of the cells should be prepared by a method that avoids virus disruption,
while allowing maximum virus release and ensuring that all cells are lysed (e.g.
three freeze/thaw cycles, followed by low-speed centrifugation). Each animal
should be inoculated subcutaneously above the scapula with a lysate obtained
from 10 7 cells. Before inoculation, it should be determined that no viable cells
are present, as development of tumours from cells would invalidate the test. The
cell lysate is suspended in PBS and inoculated in a volume of 50–100 µl into
newborn nude mice, newborn hamsters and newborn rats (2). If, at the end of
the observation period, there is no evidence of a progressively growing tumour
at the site of inoculation or at distant sites, the cell line may be considered
not to possess oncogenic activity. If tumours are observed in this assay, the
species of origin will need to be confirmed. The species of tumours that arises
in a tumorigenicity assay will be that of the cell substrate, while the species of
tumours that arises in an oncogenicity assay is that of the host (e.g. rodent). If the
cells were not lysed properly, it may be that the tumours that arose were from the
species of the cell substrate.
b. DNA
Total cellular DNA isolated from the cell substrate should be inoculated
subcutaneously above the scapulae in PBS into newborn nude mice, newborn
hamsters and newborn rats. The amount of DNA inoculated should be ≥100 µg
in 50–100 µl. Because of the concentrations necessary to achieve ≥100 µg of
DNA, it may be necessary to shear the DNA; this can be done by sonication or by
several passes in a needle and syringe. A positive-control plasmid with the test
article DNA should be inoculated into a few mice, to confirm that the cellular
DNA is not inhibitory and that the animals are susceptible to tumour induction
by DNA.
185
6. Observation period
Animals are examined weekly by observation and palpation, for evidence of
nodule formation at the site of injection. The observation period should last at
least 4 months.
7. Assessment of the inoculation site over time

(progressive or regressive growth)
If one or more nodules appear, they are measured in two perpendicular
dimensions, the measurements being recorded weekly to determine whether
the nodule grows progressively, remains stable or decreases in size over time.
Animals bearing nodules that are progressing should be killed when the nodule
reaches a size of approximately 2 cm in diameter, unless a lower limit has been
established by the authorities for the humane treatment of animals.
8. Final assessment of the inoculation site

At the end of the observation period, all animals, including the reference
group(s), are killed and examined for gross and microscopic evidence of tumour
formation at the site of injection and at other sites. Any tumour that is identified
is divided into three equal parts: (a) fixed in formalin for histopathology;
(b) used to establish a cell line, when possible; and (c) frozen for subsequent
molecular analysis.
9. Evaluation of animals for metastases

Animals are examined for microscopic evidence of metastatic lesions in sites
such as the liver, heart, lungs, spleen and regional lymph nodes.
10. Assessment of metastases (if any)

All tumours are examined to establish their relationship to the primary tumour
at the site of inoculation. If what appears to be a metastatic tumour differs
histopathologically from the primary tumour, it is necessary to consider the
possibility that this tumour developed spontaneously. This may require further
testing of the tumour itself. In such cases, appropriate follow-up studies should
be discussed and agreed with the NRA/NCL.
11. Interpretation of results

If tumours arise in the cell lysate or DNA assay, these could be induced by an
oncogenic virus or oncogenic DNA. Because of the implications for the use
of a cell substrate that contains an oncogenic agent or an oncogenic activity
for a biological, the NRA/NCL should be consulted to consider additional
186
Annex 3
experiments to identify the oncogenic agent/activity and to determine the

suitability of the use of the CCL.
Applicability
■ Cell banks: MCB or WCB taken beyond EOPC level/ECB

■ Cell types: CCL, SCL (recommended when tumorigenic cells are
used in vaccine production)
References
1. Sheng-Fowler L et al. Tumors induced in mice by direct inoculation of plasmid DNA expressing
both activated H-ras and c-myc. International Journal of Biological Sciences, 2010, 6(2):151–162.
2. Tatalick LM et al. Safety characterization of HeLa-based cell substrates used in the manufacture
of a recombinant adenoassociated virus-HIV vaccine. Vaccine, 2005, 23:2628–2638.
187

Recomendación para La Evaluacion de de Cultivo Celulares y Sustratos para El Proceso de Manufactura de Biologicos

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Recomendación para La Evaluacion de de Cultivo Celulares y Sustratos para El Proceso de Manufactura de Biologicos

Uploaded by

Copyright:

Available Formats

Annex 3

Recommendations for the evaluation of animal cell cultures

Recommendations published by WHO are intended to be scientific

EPIC-PCR exon-primed intron crossing PCR

MDCK Madin–Darby canine kidney

OIE World Organisation for Animal Health

manufacture of biological products. Although the decision-making authority lies

a prophylactic one, thus representing different risk–benefit considerations. In

The following Recommendations provide guidance to manufacturers,

product specific and are not specifically addressed;

document. Requirements or recommendations for individual products should be

encephalopathy (TSE) agents, and viruses that have been unintentionally

Biological medicinal product: biological medicinal product is a synonym

Biotherapeutic: for the purpose of this document, a biotherapeutic is a

Cells used to generate essential components of a final product (e.g. Vero

Continuous cell line (CCL): a cell line with an apparently unlimited

This definition is based on experience and the current understanding of

DNA infectivity: the capacity of cellular DNA to generate an infectious

Endogenous virus: a virus whose genome is present in an integrated

Manipulation may be as simple as the expansion of a primary cell culture

Passage: the process of transferring of cells, with or without dilution,

Population doubling: a twofold increase in cell number.

or at a distant site by metastasis.

WHO reference cell bank (RCB): a cryopreserved stock of cells prepared

conditions, such as in the vapour or liquid phase of liquid nitrogen, in aliquots

5.1.1.1 Advantages of PCCs

■ PCCs are comparatively easy to prepare using simple media and

5.1.1.2 Disadvantages of PCCs

■ Contamination by infectious agents is a higher risk than with DCLs

5.1.2 Diploid cell lines (DCLs)

established as cell lines with apparently stable characteristics over

Substantial experience has been accumulated on the cytogenetics of

5.1.2.1 Advantages of DCLs

■ DCLs can be well characterized and standardized.

5.1.2.2 Disadvantages of DCLs

■ DCLs are not easy to use in large-scale production, although they

5.1.3 Continuous cell lines (CCLs)

5.1.3.1 Advantages of CCLs

■ CCLs can be characterized extensively and their culture conditions

5.1.3.2 Disadvantages of CCLs

■ CCLs may express endogenous viruses, and some are tumorigenic in

5.1.4 Stem cell lines (SCLs)

5.1.4.1 Advantages of SCLs

■ SCLs can be well characterized and their culture conditions

5.1.4.2 Disadvantages of SCLs

■ Subculture techniques commonly used for SCLs are laborious.

5.2 Potential risks and risk mitigation associated with

5.2.1 Viruses and other transmissible agents

potentially zoonotic viruses. While contamination with the rodent viruses in

There may be as yet undiscovered microbial agents for which there is

5.2.2 Cellular DNA

oncogenic risk. The polyoma virus genome is infectious in mice at about 50 pg

5.2.3 Cellular RNA

5.2.4 Growth-promoting proteins

dependent upon continued administration of the growth factor. Therefore, the

Part A. General recommendations applicable to

A.2 Principles of good cell culture practice

■ authenticity, including identity, provenance and genotypic/phenotypic

■ absence of microbiological contamination;

A.2.2 Manipulation of cell cultures

A.2.2.1 Detachment and subculture

A.2.2.3 Introduction of contamination